|Publication number||US6697935 B1|
|Application number||US 08/958,716|
|Publication date||24 Feb 2004|
|Filing date||23 Oct 1997|
|Priority date||23 Oct 1997|
|Also published as||CA2299348A1, CA2299348C, CN1112636C, CN1276888A, EP1029269A1, EP1029269B1, WO1999021081A1|
|Publication number||08958716, 958716, US 6697935 B1, US 6697935B1, US-B1-6697935, US6697935 B1, US6697935B1|
|Inventors||John Michael Borkenhagen, Richard James Eickemeyer, William Thomas Flynn, Andrew Henry Wottreng|
|Original Assignee||International Business Machines Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (69), Non-Patent Citations (79), Referenced by (239), Classifications (12), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
“The present invention relates to the following U.S. applications, the subject matter of which is hereby incorporated by reference: (1) U.S. Ser. No. 08/957,002 entitled Thread Switch Control in a Multithreaded Processor System, filed concurrently herewith; (2) U.S. Pat. No. 6,105,051 entitled Apparatus and Method to Guarantee Forward progress in Execution of Threads in a Multithreaded Processor, filed concurrently herewith; (3) U.S. Pat. No. 6,212,544 entitled Altering Thread Priorities in a Multithreaded Processor, filed concurrently herewith; (4) U.S. Pat. No. 6,076,157 entitled Method and Apparatus to Force a Thread Switch in a Multithreaded Processor, filed concurrently herewith; (5) U.S. Pat. No. 6,088,788 entitled Background Completion of Instruction and Associated Fetch Request in a Multithread Processor (6) U.S. Pat. No. 6,000,011 entitled Multi-Entry Fully Associative Transition Cache; (7) U.S. Pat. No. 6,000,012 entitled Method and Apparatus for Prioritizing and Routing Commands from a Command Source to a Command Sink; (8) U.S. Pat. No. 6,035,424 entitled Method and Apparatus for Tracking Processing of a Command; (9) U.S. Pat. No. 6,049,867 entitled Method and System for Multithread Switching Only When a Cache Miss Occurs at a Second or Higher Lever; and (10) U.S. Pat. No. 5,778,243 entitled Multithreaded Cell for a Memory.”
The present invention relates in general to an improved method for and apparatus of a computer data processing system; and in particular, to an improved high performance multithreaded computer data processing system and method embodied in the hardware of the processor.
The fundamental structure of a modern computer includes peripheral devices to communicate information to and from the outside world; such peripheral devices may be keyboards, monitors, tape drives, communication lines coupled to a network, etc. Also included in the basic structure of the computer is the hardware necessary to receive, process, and deliver this information from and to the outside world, including busses, memory units, input/output (I/O) controllers, storage devices, and at least one central processing unit (CPU), etc. The CPU is the brain of the system. It executes the instructions which comprise a computer program and directs the operation of the other system components.
From the standpoint of the computer's hardware, most systems operate in fundamentally the same manner. Processors actually perform very simple operations quickly, such as arithmetic, logical comparisons, and movement of data from one location to another. Programs which direct a computer to perform massive numbers of these simple operations give the illusion that the computer is doing something sophisticated. What is perceived by the user as a new or improved capability of a computer system, however, may actually be the machine performing the same simple operations, but much faster. Therefore continuing improvements to computer systems require that these systems be made ever faster.
One measurement of the overall speed of a computer system, also called the throughput, is measured as the number of operations performed per unit of time. Conceptually, the simplest of all possible improvements to system speed is to increase the clock speeds of the various components, particularly the clock speed,of the processor. So that if everything runs twice as fast but otherwise works in exactly the same manner, the system will perform a given task in half the time. Computer processors which were constructed from discrete components years ago performed significantly faster by shrinking the size and reducing the number of components; eventually the entire processor was packaged as an integrated circuit on a single chip. The reduced size made it possible to increase the clock speed of the processor, and accordingly increase system speed.
Despite the enormous improvement in speed obtained from integrated circuitry, the demand for ever faster computer systems still exists. Hardware designers have been able to obtain still further improvements in speed by greater integration, by further reducing the size of the circuits, and by other techniques. Designer, however, think that physical size reductions cannot continue indefinitely and there are limits to continually increasing processor clock speeds. Attention has therefore been directed to other approaches for further improvements in overall speed of the computer system.
Without changing the clock speed, it is still possible to improve system speed by using multiple processors. The modest cost of individual processors packaged on integrated circuit chips has made this practical. The use of slave processors considerably improves system speed by off-loading work from the CPU to the slave processor. For instance, slave processors routinely execute repetitive and single special purpose programs, such as input/output device communications and control. It is also possible for multiple CPUs to be placed in a single computer system, typically a host-based system which services multiple users simultaneously. Each of the different CPUs can separately execute a different task on behalf of a different user, thus increasing the overall speed of the system to execute multiple tasks simultaneously. It is much more difficult, however, to improve the speed at which a single task, such as an application program, executes. Coordinating the execution and delivery of results of various functions among multiple CPUs is a tricky business. For slave I/O processors this is not so difficult because the functions are pre-defined and limited but for multiple CPUs executing general purpose application programs it is much more difficult to coordinate functions because, in part, system designers do not know the details of the programs in advance, Most application programs follow a single path or flow of steps performed by the processor. While it is sometimes possible to break up this single path into multiple parallel paths, a universal application for doing so is still being researched. Generally, breaking a lengthy task into smaller tasks for parallel processing by multiple processors is done by a software engineer writing code on a case-by-case basis. This ad hoc approach is especially problematic for executing commercial transactions which are not necessarily repetitive or predictable.
Thus, while multiple processors improve overall system performance, there are still many reasons to improve the speed of the individual CPU. If the CPU clock speed is given, it is possible to further increase the speed of the CPU, i.e., the number of operations executed per second, by increasing the average number of operations executed per clock cycle. A common architecture for high performance, single-chip microprocessors is the reduced instruction set computer (RISC) architecture characterized by a small simplified set of frequently used instructions for rapid execution, those simple operations performed quickly mentioned earlier. As semiconductor technology has advanced, the goal of RISC architecture has been to develop processors capable of executing one or more instructions on each clock cycle of the machine. Another approach to increase the average number of operations executed per clock cycle is to modify the hardware within the CPU. This throughput measure, clock cycles per instruction, is commonly used to characterize architectures for high performance processors. Instruction pipelining and cache memories are computer architectural features that have made this achievement possible. Pipeline instruction execution allows subsequent instructions to begin execution before previously issued instructions have finished. Cache memories store frequently used and other data nearer the processor and allow instruction execution to continue, in most cases, without waiting the full access time of a main memory. Some improvement has also been demonstrated with multiple execution units with look ahead hardware for finding instructions to execute in parallel.
The performance of a conventional RISC processor can be further increased in the superscalar computer and the Very Long Instruction Word (VLIW) computer, both of which execute more than one instruction in parallel per processor cycle. In these architectures, multiple functional or execution units are provided to run multiple pipelines in parallel. In a superscalar architecture, instructions may be completed in-order and out-of-order. In-order completion means no instruction can complete before all instructions dispatched ahead of it have been completed. Out-of-order completion means that an instruction is allowed to complete before all instructions ahead of it have been completed, as long as a predefined rules are satisfied.
For both in-order and out-of-order execution in superscalar systems, pipelines will stall under certain circumstances. An instruction that is dependent upon the results of a previously dispatch ed instruction that has not yet completed may cause the pipeline to stall. For instance, instructions dependent on a load/store instruction in which the necessary data is not in the cache, i.e., a cache miss, cannot be executed until the data becomes available in the cache. Maintaining the requisite data in the cache necessary for continued execution and to sustain a high hit ratio, i.e., the number of requests for data compared to the number of times the data was readily available in the cache, is not trivial especially for computations involving large data structures. A cache miss can cause the pipelines to stall for several cycles, and the total amount of memory latency will be severe if the data is not available most of the time. Although memory devices used for main memory are becoming faster, the speed gap between such memory chips and high-end processors is becoming increasingly larger. Accordingly, a significant amount of execution time in current high-end processor designs is spent waiting for resolution of cache misses and these memory access delays use an increasing proportion of processor execution time.
And yet another technique to improve the efficiency of hardware within the CPU is to divide a processing task into independently executable sequences of instructions called threads. This technique is related to breaking a larger task into smaller tasks for independent execution by different processors except here the threads are to be executed by the same processor. When a CPU then, for any of a number of reasons, cannot continue the processing or execution of one of these threads, the CPU switches to and executes another thread. This is the subject of the invention described herein which incorporates hardware multithreading to tolerate memory latency. The term “multithreading” as defined in the computer architecture community is not the same as the software use of the term which means one task subdivided into multiple related threads. In the architecture definition, the threads may be independent. Therefore “hardware multithreading” is often used to distinguish the two uses of the term. The present invention incorporates the term multithreading to connote hardware multithreading.
Multithreading permits the processors' pipeline(s) to do useful work on different threads when a pipeline stall condition is detected for the current thread. Multithreading also permits processors implementing non-pipeline architectures to do useful work for a separate thread when a stall condition is detected for a current thread. There are two basic forms of multithreading. A traditional form is to keep N threads, or states, in the processor and interleave the threads on a cycle-by-cycle basis. This eliminates all pipeline dependencies because instructions in a single thread are separated. The other form of multithreading, and the one considered by the present invention, is to interleave the threads on some long-latency event.
Traditional forms of multithreading involves replicating the processor registers for each thread. For instance, for a processor implementing the architecture sold under the trade name PowerPC™ to perform multithreading, the processor must maintain N states to run N threads. Accordingly, the following are replicated N times: general purpose registers, floating point registers, condition registers, floating point status and control register, count register, link register, exception register, save/restore registers, and special purpose registers. Additionally, the special buffers, such as a segment lookaside buffer, can be replicated or each entry can be tagged with the thread number and, if not, must be flushed on every thread switch. Also, some branch prediction mechanisms, e.g., the correlation register and the return stack, should also be replicated. Fortunately, there is no need to replicate some of the larger functions of the processor such as: level one instruction cache (L1 I-cache), level one data cache (L1 D-cache), instruction buffer, store queue, instruction dispatcher, functional or execution units, pipelines, translation lookaside buffer (TLB), and branch history table. When one thread encounters a delay, the processor rapidly switches to another thread. The execution of this thread overlaps with the memory delay on the first thread.
Existing multithreading techniques describe switching threads on a cache miss or a memory reference. A primary example of this technique may be reviewed in “Sparcle: An Evolutionary Design for Large-Scale Multiprocessors,” by Agarwal et al., IEEE Micro Volume 13, No. 3, pp. 48-60, June 1993. As applied in a RISC architecture, multiple register sets normally utilized to support function calls are modified to maintain multiple threads. Eight overlapping register windows are modified to become four non-overlapping register sets, wherein each register set is a reserve for trap and message handling. This system discloses a thread switch which occurs on each first level cache miss that results in a remote memory request. While this system represents an advance in the art, modern processor designs often utilize a multiple level cache or high speed memory which is attached to the processor. The processor system utilizes some well-known algorithm to decide what portion of its main memory store will be loaded within each level of cache and thus, each time a memory reference occurs which is not present within the first level of cache the processor must attempt to obtain that memory reference from a second or higher level of cache.
It should thus be apparent that a need exists for an improved data processing system which can reduce delays due to memory latency in a multilevel cache system utilized in conjunction with a multithread data processing system.
An object of the present invention is to provide an improved data processing system and method for multithreaded processing embodied in the hardware of the processor. This object is achieved by a multithreaded processor capable of switching execution between two threads of instructions, and thread switch logic embodied in hardware registers with optional software override of thread switch conditions. An added advantage of the thread switch logic is that processing of various threads of instructions allows optimization of the use of the processor among the threads.
Another object of the present invention is to improve multithreaded computer processing by allowing the processor to execute a second thread of instructions thereby increasing processor utilization which is otherwise idle because it is retrieving necessary data and/or instructions from various memory elements, such as caches, memories, external I/O, direct access storage device for a first thread.
An additional object of the present invention is to provide a multithread data processing system and method which performs conditional thread switching wherein the conditions of thread switching can be different per thread or can be changed during processing by the use of a software thread control manager.
It is yet another object of the invention to provide a hardware register containing bits embodying the events which can cause a multithreaded processor to switch threads. The feature of this invention is that the bits in this hardware register, the thread switch control register, can be enabled. This hardware register has the further advantage of improving processor performance because it is much faster than software thread switch control.
These and other related objects are achieved by providing a computer system having a multithreaded processor capable of switching processing between at least two threads of instructions when the multithreaded processor experiences one of a plurality of processor latency events. The computer system also has at least one thread state register operatively connected to the multithreaded processor, to store a state of the threads of instructions wherein the state of each thread of instructions changes when the processor switches processing to each thread. The system also has at least one thread switch control register operatively connected to the thread state register(s) and to the multithreaded processor, to store a plurality of thread switch control events which thread switch control events are enabled by setting a corresponding plurality of enable bits. The computer system further comprises a plurality of internal connections connecting the multithreaded processor to a plurality of memory elements. Access to any of the,memory elements by the multithreaded processor causes a processor latency event and the invention also has at least one external connection connecting the multithreaded processor to an external memory device, a communication device, a computer network, or an input/output device wherein access to any of the devices or the network by the multithreaded processor also causes a plurality of processor latency events. When one of the threads executing in the multithreaded processor is unable to continue execution because of one of the processor latency events and when that processor latency event is a thread switch control event whose bit is enabled, the multithreaded processor switches execution to another of the threads.
The thread switch control register has a plurality of bits, each associated uniquely with one of a plurality of thread switch control events and if one of the bits is enabled, the thread switch control event associated with that bit causes the multithreaded processor(s) to switch from one thread of instructions to another thread of instructions. The thread switch control register is programmable. Moreover, the enablement of a particular bit can be dynamically changed by either operating software or by an instruction in one of the threads.
The computer processing system may have more than one thread switch control register wherein the bit values of one thread switch control register differs from the bit values of another of said thread switch control registers.
Typically, there can be many thread switch control events, for instance, a data miss from at least one of the following: a L1-data cache, a L2 cache, storage of data that crosses a double word boundary, or an instruction miss from at least one of the following: a L1-instruction cache, a translation lookaside buffer, or a data and/or instruction miss from main memory, or an error in address translation of data and/or an instruction. Access to an I/O device external to the processor or to another processor may also be thread switch control events. Other thread switch control events comprise a forward progress count of a number of times said one of a plurality of threads has been switched from a one multithreaded processor with no instruction of the one of a plurality of threads executing, and a time-out period in which no useful work was done by the at least one processor.
The computer processing system of the invention comprises means for processing a plurality of threads of instructions; means for indicating when the processing means stalls because one of the threads experiences a processor latency event; means for registering a plurality of thread switch control events; and means for determining if the processor latency event is one of the plurality of thread switch control events. The processing system may also comprise means for enabling the processing means to switch processing to another thread if the processor latency event is a thread switch control event.
The invention is also a method to determine the contents of a thread switch control register, comprising the steps of counting a first number of processor cycles in which a multithreaded processor is stalled because of processor latency event and counting a second number of processor cycles required for the multithreaded processor to switch processing of a first thread of instructions to a second thread of instructions, then assigning the processor latency event to be a thread switch control event by setting an enable bit in the thread switch control register if the first number is greater than the second number. Then if the enable bit is enabled, the method comprises outputting a signal to switch threads when the multithreaded processor experiences the thread switch control event if the enable bit is enabled.
The method of computer processing of the invention also comprises the steps of storing a state of a thread in a thread state register and storing a plurality of thread switch control events in a thread switch control register. Then when the state of the thread changes, a signal is output to the thread state register and the changed state of the thread is compared with the plurality of thread switch control events. If the changed state results from a thread switch control event, a signal is output to a multithreaded processor to switch execution from the thread.
Other objects, features and characteristics of the present invention; methods, operation, and functions of the related elements of the structure; combination of parts; and economies of manufacture will become apparent from the following detailed description of the preferred embodiments and accompanying drawings, all of which form a part of this specification, wherein like reference numerals designate corresponding parts in the various figures.
The invention itself, however, as well as a preferred mode of use, further objectives and advantages thereof, will best be understood by reference to the following detailed description of an illustrative embodiment when read in conjunction with the accompanying drawings, wherein:
FIG. 1 is a block diagram of a computer system capable of implementing the invention described herein.
FIG. 2 illustrates a high level block diagram of a multithreaded data processing system according to the present invention.
FIG. 3 illustrates a block diagram of the storage control unit of FIG. 2.
FIG. 4 illustrates a block diagram of the thread switch logic, the storage control unit and the instruction unit of FIG. 2.
FIG. 5 illustrate the changes of state of a thread as the thread experiences different thread switch events shown in FIG. 4.
FIG. 6 is a flow chart of the forward progress count of the invention.
With reference now to the figures and in particular with reference to FIG. 1, there is depicted a high level block diagram of a computer data processing system 10 which may be utilized to implement the method and system of the present invention. The primary hardware components and interconnections of a computer data processing system 10 capable of utilizing the present invention are shown in FIG. 1. Central processing unit (CPU) 100 for processing instructions is coupled to caches 120, 130, and 150. Instruction cache 150 stores instructions for execution by CPU 100. Data caches 120, 130 store data to be used by CPU 100. The caches communicate with random access memory in main memory 140. CPU 100 and main memory 140 also communicate via bus interface 152 with system bus 155. Various input/output processors (IOPs) 160-168 attach to system bus 155 and support communication with a variety of storage and input/output (I/O) devices, such as direct access storage devices (DASD) 170, tape drives 172, remote communication lines 174, workstations 176, and printers 178. It should be understood that FIG. 1 is intended to depict representative components of a computer data processing system 10 at a high level, and that the number and types of such components may vary.
Within the CPU 100, a processor core 110 contains specialized functional units, each of which perform primitive operations, such as sequencing instructions, executing operations involving integers, executing operations involving real numbers, transferring values between addressable storage and logical register arrays. FIG. 2. illustrates a processor core 100. In a preferred embodiment, the processor core 100 of the data processing system 10 is a single integrated circuit, pipelined, superscalar microprocessor, which may be implemented utilizing any computer architecture such as the family of RISC processors sold under the trade name PowerPC™; for example, the PowerPC™ 604 microprocessor chip sold by IBM.
As will be discussed below, the data processing system 10 preferably includes various units, registers, buffers, memories, and other sections which, are all preferably formed by integrated circuitry. It should be understood that in the figures, the various data paths have been simplified; in reality, there are many separate and parallel data paths into and out of the various components. In addition, various components not germane to the invention described herein have been omitted, but it is to be understood that processors contain additional units for additional functions. The data processing system 10 can operate according to reduced instruction set computing, RISC, techniques or other computing techniques.
As represented in FIG. 2, the processor core 100 of the data processing system 10 preferably includes a level one data cache, L1 D-cache 120, a level two L2 cache 130, a main memory 140, and a level one instruction cache, L1 I-cache 150, all of which are operationally interconnected utilizing various bus connections to a storage control unit 200. As shown in FIG. 1, the storage control unit 200 includes a transition cache. 210 for interconnecting the L1 D-cache 120 and the L2 cache 130, the main memory 140, and a plurality of execution units. The L1 D-cache 120 and L1 I-cache 150 preferably are provided on chip as part of the processor 100 while the main memory 140 and the L2 cache 130 are provided off chip. Memory system 140 is intended to represent random access main memory which may or may not be within the processor core 100 and, and other data buffers and caches, if any, external to the processor core 100, and other external memory, for example, DASD 170, tape drives 172, and workstations 176, shown in FIG. 1. The L2 cache 130 is preferably a higher speed memory system than the main memory 140, and by storing selected data within the L2 cache 130, the memory latency which occurs as a result of a reference to the main memory 140 can be minimized. As shown in FIG. 1, the L2 cache 130 and the main memory 140 are directly connected to both the L1 I-cache 150 and an instruction unit 220 via the storage control unit 200.
Instructions from the L1 I-cache 150 are preferably output to an instruction unit 220 which, in accordance with the method and system of the present invention, controls the execution of multiple threads by the various subprocessor units, e.g., branch unit 260, fixed point unit 270, storage control unit 200, and floating point unit 280 and others as specified by the architecture of the data processing system 10. In addition to the various execution units depicted within FIG. 1, those skilled in the art will appreciate that modem superscalar microprocessor systems often include multiple versions of each such execution unit which may be added without departing from the spirit and scope of the present invention. Most of these units will have as an input source operand information from various registers such as general purpose registers GPRs 272, and floating point registers FPRs 282. Additionally, multiple special purpose register SPRs 274 may be utilized. As shown in FIG. 2, the storage control unit 200 and the transition cache 210 are directly connected to general purpose registers 272 and the floating point registers 282. The general purpose registers 272 are connected to the special purpose registers 274.
Among the functional hardware units unique to this multithreaded processor 100 is the thread switch logic 400 and the transition cache 210. The thread switch logic 400 contains various registers that determine which thread will be the active or the executing thread. Thread switch logic 400 is operationally connected to the storage control unit 200, the execution units 260, 270, and 280, and the instruction unit 220. The transition cache 210 within the storage control unit 200 must be capable of implementing multithreading. Preferably, the storage control unit 200 and the transition cache 210 permit at least one outstanding data request per thread. Thus, when a first thread is suspended in response to, for example, the occurrence of L1 D-cache miss, a second thread would be able to access the L1 D-cache 120 for data present therein. If the second thread also results in L1 D-cache miss, another data request will be issued and thus multiple data requests must be maintained within the storage control unit 200 and the transition cache 210. Preferably, transition cache 210 is the transition cache of U.S. Pat. No. 6,000,011 entitled Multi-Entry Fully Associative Transition Cache, hereby incorporated by reference. The storage control unit 200, the execution units 260, 270, and 280 and the instruction unit 220 are all operationally connected to the thread switch logic 400 which determines which thread to execute.
As illustrated in FIG. 2, a bus 205 is provided between the storage control unit 200 and the instruction unit 220 for communication of, e.g., data requests to the storage control unit 200, and a L2 cache 130 miss to the instruction unit 22,0. Further, a translation lookaside buffer TLB 250 is provided which contains virtual-to-real address mapping. Although not illustrated within the present invention various additional high level memory mapping buffers may be provided such as a segment lookaside buffer which will operate in a manner similar to the translation lookaside buffer 250.
FIG. 3 illustrates the storage control unit 200 in greater detail, and, as the name implies, this unit controls the input and output of data and instructions from the various storage units, which include the various caches, buffers and main memory. As shown in FIG. 3, the storage control unit 200 includes the transition cache 210 functionally connected to the L1 D-cache 120, multiplexer 360, the L2 cache 130, and main memory 140. Furthermore, the transition cache 210 receives control signals from sequencers 350. The sequencers 350 include a plurality of sequencers, preferably three, for handling instruction and/or data fetch requests. Sequencers 350 also output control signals to the transition cache 210, the L2 cache 130, as well as receiving and transmitting control signals to and from the main memory 140.
Multiplexer 360 in the storage control unit 200 shown in FIG. 3 receives data from the L1 D-cache 120, the transition cache 210, the L2 cache 130, main memory 140, and, if data is to be stored to memory, the execution units 270 and 280. Data from one of these sources is selected by the multiplexer 360 and is output to the L1 D-cache 120 or the execution units in response to a selection control signal received from the sequencers 350. Furthermore, as shown in FIG. 3, the sequencers 350 output a selection signal to control a second multiplexer 370. Based on this selection signal from the sequencers 350, the multiplexer 370 outputs the data from the L2 cache 130 or the main memory 140 to the L1 I-cache 150 or the instruction unit 220. In producing the above-discussed control and selection signals, the sequencers 350 access and update the L1 directory 320 for the L1 D-cache 120 and the L2 directory 330 for the L2 cache 130.
With respect to the multithreading capability of the processor described herein, sequencers 350 of the storage control unit 200 also output signals to thread switch logic 400 which indicate the state of data and instruction requests. So, feedback from the caches 120, 130 and 150, main memory 140, and the translation lookaside buffer 250 is routed to the sequencers 350 and is then communicated to thread switch logic 400 which may result in a thread switch, as discussed below. Note that any device wherein an event designed to cause a thread switch in a multithreaded processor occurs will be operationally connected to sequencers 350.
FIG. 4 is a logical representation and block diagram of the thread switch logic hardware 400 that determines whether a thread will be switched and, if so, what thread. Storage control unit 200 and instruction unit 220 are interconnected with thread switch logic 400. Thread switch logic 400 preferably is incorporated into the instruction unit 220 but if there are many threads the complexity of the thread switch logic 400 may increase so that the logic is external to the instruction unit 220. For ease of explanation, thread switch logic 400 is illustrated external to the instruction unit 220.
Some events which result in a thread to be switched in this embodiment are communicated on lines 470, 472, 474, 476, 478, 480, 482, 484, and 486 from the sequencers 350 of the storage control unit 200 to the thread switch logic 400. Other latency events can cause thread switching; this list is not intended to be inclusive; rather it is only representative of how the thread switching can be implemented. A request for an instruction by either the first thread T0 or the second thread T1 which is not in the instruction unit 220 is an event which can result in a thread switch, noted by 470 and 472 in FIG. 4, respectively. Line 474 indicates when the active thread, whether T0 or T1, experiences a L1 D-cache 120 miss. Cache misses of the L2 cache 130 for either thread T0 or T1 is noted at lines 476 and 478, respectively. Lines 480 and 482 are activated when data is returned for continued execution of the T0 thread or for the T1 thread, respectively. Translation lookaside buffer misses and completion of a table walk are indicated by lines 484 and 486, respectively.
These events are all fed into the thread switch logic 400 and more particularly to the thread state registers 440 and the thread switch controller 450. Thread switch logic 400 has one thread state register for each thread. In the embodiment described herein, two threads are represented so there is a T0 state register 442 for a first thread T0 and a T1 state register 444 for a second thread T1, to be described herein. Thread switch logic 400 comprises a thread switch control register 410 which controls what events will result in a thread switch. For instance, the thread switch control register 410 can block events that cause state changes from being seen by the thread switch controller 450 so that a thread may not be switched as a result of a blocked event. The thread state registers and the logic of changing threads are the subject of a U.S. patent application Ser. No. 08/957,002, filed concurrently and herein incorporated by reference. The forward progress count register 420 is used to prevent thrashing and may be included in the thread switch control register 410. The forward progress count register 420 is the subject of U.S. Pat. No. 6,105,051, filed concurrently and herein incorporated by reference. Thread switch time-out register 430, the subject of U.S. Pat. No. 6,076,157 filed concurrently and herein incorporated by reference, allocates fairness and livelock issues. Also, thread priorities can be altered using software 460, the subject of U.S. Pat. No. 6,212,544 filed concurrently and herein incorporated by reference. Finally, but not to be limitative, the thread switch controller 450 comprises a myriad of logic gates which represents the culmination of all logic which actually determines whether a thread is switched, what thread, and under what circumstances. Each of these logic components and their functions are set forth in further detail.
Thread State Registers
Thread state registers 440 comprise a state register for each thread and, as the name suggests, store the state of the corresponding thread; in this case, a T0 thread state register 442 and a T1 thread state register 444. The number of bits and the allocation of particular bits to describe the state of each thread can be customized for a particular architecture and thread switch priority scheme. An example of the allocation of bits in the thread state registers 442, 444 for a multithreaded processor having two threads is set forth in the table below.
Thread State Register Bit Allocation
0 = Instruction
1 = Data
Miss type sequencer
00 = None
01 = Translation lookaside buffer miss (check bit 0 for I/D)
10 = L1 cache miss
11 = L2 cache miss
0 = Transition to current state does not result in thread switch
1 = Transition to current state results in thread switch
0 = Load
1 = Store
Forward progress counter
111 = Reset (instruction has completed during this thread)
000 = 1st execution of this thread w/o instruction complete
001 = 2nd execution of this thread w/o instruction complete
010 = 3rd execution of this thread w/o instruction complete
011 = 4th execution of this thread w/o instruction complete
100 = 5th execution of this thread w/o instruction complete
Priority (could be set by software)
00 = Medium
01 = Low
10 = High
11 = <Illegal>
Reserved if 64 bit implementation
In the embodiment described herein, bit 0 identifies whether the miss or the reason the processor stalled execution is a result of a request for an instruction or for data. Bits 1 and 2 indicate if the requested information was not available and if so, from what hardware, i.e., whether the translated address of the data or instruction was not in the translation lookaside buffer 250, or the data or instruction itself was not in the L1 D-cache 120 or the L2 cache 130, as further explained in the description of FIG. 5. Bit 3 indicates whether the change of state of a thread results in a thread switch. A thread may change state without resulting in a thread switch. For instance, if a thread switch occurs when thread T1 experiences an L1 cache miss, then if thread T1 experiences a L2 cache miss, there will be no thread switch because the thread already switched on a L1 cache miss. The state of T1, however, still changes. Alternatively, if by choice, the thread switch logic 400 is configured or programmed not to switch on a L1 cache miss, then when a thread does experience an L1 cache miss, there will be no thread switch even though the thread changes state. Bit 8 of the thread state registers 442 and 444 is assigned to whether the information requested by a particular thread is to be loaded into the processor core or stored from the processor core into cache or main memory. Bits 15 through 17 are allocated to prevent thrashing, as discussed later with reference to the forward progress count register 420. Bits 18 and 19 can be set in the hardware or could be set by software to indicate the priority of the thread.
FIG. 5 represents four states in the present embodiment of a thread processed by the data processing system 10 and these states are stored in the thread state registers 440, bit positions 1:2. State 00 represents the “ready” state, i.e., the thread is ready for processing because all data and instructions required are available; state 10 represents the thread state wherein the execution of the thread within the processor is stalled because the thread is waiting for return of data into either the L1 D-cache 120 or the return of an instruction into the L1 I-cache 150; state 11 represents that the thread is waiting for return of data into the L2 cache 130; and the state 01 indicates that there is a miss on the translation lookaside buffer 250, i.e., the virtual address was in error or wasn't available, called a table walk. Also shown in FIG. 5 is the hierarchy of thread states wherein state 00, which indicates the thread is ready for execution, has the highest priority. Short latency events are preferably assigned a higher priority.
FIG. 5 also illustrates the change of states when data is retrieved from various sources. The normal uninterrupted execution of a thread T0 is represented in block 510 as state 00. If a L1 D-cache or I-cache miss occurs, the thread state changes to state 10, as represented in block 512, pursuant to a signal sent on line 474 (FIG. 4) from the storage control unit 200 or line 470 (FIG. 4) from the instruction unit 220, respectively. If the required data or instruction is in the L2 cache 130 and is retrieved, then normal execution of T0 resumes at block 510. Similarly block 514 of FIG. 5 represents a L2 cache miss which changes the state of thread of either T0 or T1 to state 11 when storage control unit 200 signals the miss on lines 476 or 478. (FIG. 4). When the instructions or data in the L2 cache are retrieved from main memory 140 and loaded into the processor core 100 as indicated on lines 480 and 482 (FIG. 4), the state again changes back to state 00 at block 510. The storage control unit 200 communicates to the thread registers 440 on line 484 (FIG. 4) when the virtual address for requested information is not available in the translation lookaside buffer 250, indicated as block 516, as a TLB miss or state 01. When the address does become available or if there is a data storage interrupt instruction as signaled by the storage control unit 200 on line 486 (FIG. 4), the state of the thread then returns to state 00, meaning ready for execution.
The number of states, and what each state represents is freely selectable by the computer architect. For instance, if a thread has multiple L1 cache misses, such as both a L1 I-cache miss and L1 D-cache miss, a separate state can be assigned to each type of cache miss. Alternatively, a single thread state could be assigned to represent more than one event or occurrence.
An example of a thread switch algorithm for two threads of equal priority which determines whether to switch threads is given. The algorithm can be expanded and modified accordingly for more threads and thread switch conditions according to the teachings of the invention. The interactions between the state of each thread stored in the thread state registers 440 (FIG. 4) and the priority of each thread by the thread switching algorithm are dynamically interrogated each cycle. If the active thread T0 has a L1 miss, the algorithm will cause a thread switch to the dormant thread T1 unless the dormant thread T1 is waiting for resolution of a L2 miss. If a switch did not occur and the L1 cache miss of active thread T0 turns into a L2 cache miss, the algorithm then directs the processor to switch to the dormant thread T1 regardless of the T1's state. If both threads are waiting for resolution of a L2 cache miss, the thread first having the L2 miss being resolved becomes the active thread. At every switch decision time, the action taken is optimized for the most likely case, resulting in the best performance. Note that thread switches resulting from a L2 cache miss are conditional on the state of the other thread, if not extra thread switches would occur resulting in loss of performance.
Thread Switch Control Register
In any multithreaded processor, there are latency and performance penalties associated with switching threads. In the multithreaded processor in the preferred embodiment described herein, this latency includes the time required to complete execution of the current thread to a point where it can be interrupted and correctly restarted when it is next invoked, the time required to switch the thread-specific hardware facilities from the current thread's state to the new thread's state, and the time required to restart the new thread and begin its execution. Preferably the thread-specific hardware facilities operable with the invention include the thread state registers described above and the memory cells described in U.S. Pat. No. 6,778,243 entitled Multithreaded Storage Cell, herein incorporated by reference. In order to achieve optimal performance in a coarse grained multithreaded data processing system, the latency of an event which generates a thread switch must be greater than the performance cost associated with switching threads in a multithreaded mode, as opposed to the normal single-threaded mode.
The latency of an event used to generate a thread switch is dependent upon both hardware and software. For example, specific hardware considerations in a multithreaded processor include the speed of external SRAMs used to implement an L2 cache external to the processor chip. Fast SRAMs in the L2 cache reduce the average latency of an L1 miss while slower SRAMS increase the average latency of an L1 miss. Thus, performance is gained if one thread switch event is defined as a L1 cache miss in hardware having an external L2 cache data access latency greater than the thread switch penalty. As an example of how specific software code characteristics affect the latency of thread switch events, consider the L2 cache hit-to-miss ratio of the code, i.e., the number of times data is actually available in the L2 cache compared to the number of times data must be retrieved from main memory because data is not in the L2 cache. A high L2 hit-to-miss ratio reduces the average latency of an L1 cache miss because the L1 cache miss seldom results in a longer latency L2 miss. A low L2 hit-to-miss ratio increases the average latency of an L1 miss because more L1 misses result in longer latency L2 misses. Thus, a L1 cache miss could be disabled as a thread switch event if the executing code has a high L2 hit-to-miss ratio because the L2 cache data access latency is less than the thread switch penalty. A L1 cache miss would be enabled as a thread switch event when executing software code with a low L2 hit-to-miss ratio because the L1 cache miss is likely to turn into a longer latency L2 cache miss.
Some types of latency events are not readily detectable. For instance, in some systems the L2 cache outputs a signal to the instruction unit when a cache miss occurs. Other L2 caches, however, do not output such a signal, as in for example, if the L2 cache controller were on a separate chip from the processor and accordingly, the processor cannot readily determine a state change. In these architectures, the processor can include a cycle counter for each outstanding L1 cache miss. If the miss data has not been returned from the L2 cache after a predetermined number of cycles, the processor acts as if there had been a L2 cache miss and changes the thread's state accordingly. This algorithm is also applicable to other cases where there are more than one distinct type of latency. As an example only, for a L2 cache miss in a multiprocessor, the latency of data from main memory may be significantly different than the latency of data from another processor. These two events may be assigned different states in the thread state register. If no signal exists to distinguish the states, a counter may be used to estimate which state the thread should be in after it encounters a L2 cache miss.
The thread switch control register 410 is a software programmable register which selects the events to generate thread switching and has a separate enable bit for each defined thread switch control event. Although the embodiment described herein does not implement a separate thread switch control register 410 for each thread, separate thread switch control registers 410 for each thread could be implemented to provide more flexibility and performance at the cost of more hardware and complexity. Moreover, the thread switch control events in one thread switch control register need not be identical to the thread switch control events in any other thread switch control register.
The thread switch control register 410 can be written by a service processor with software such as a dynamic scan communications interface disclosed in U.S. Pat. No. 5,079,725 entitled Chip Identification Method for Use with Scan Design Systems and Scan Testing Techniques or by the processor itself with software system code. The contents of the thread switch control register 410 is used by the thread switch controller 450 to enable or disable the generation of a thread switch. A value of one in the register 410 enables the thread switch control event associated with that bit to generate a thread switch. A value of zero in the thread switch control register 410 disables the thread switch control event associated with that bit from generating a thread switch. Of course, an instruction in the executing thread could disable any or all of the thread switch conditions for that particular or for other threads. The following table shows the association between thread switch events and their enable bits in the register 410.
Thread Switch Control Register Bit Assignment
Switch on L1 data cache fetch miss
Switch on L1 data cache store miss
Switch on L1 instruction cache miss
Switch on instruction TLB miss
Switch L2 cache fetch miss
Switch on L2 cache store miss
Switch on L2 instruction cache miss
Switch on data TLB/segment lookaside buffer miss
Switch on L2 cache miss and dormant thread not
L2 cache miss
Switch when thread switch time-out value reached
Switch when L2 cache data returned
Switch on IO external accesses
Switch on double-X store: miss on first of two*
Switch on double-X store: miss on second of two*
Switch on store multiple/string: miss on any access
Switch on load multiple/string: miss on any access
Switch on double-X load: miss on first of two*
Switch on double-X load: miss on second of two*
Switch on or 1,1,1 instruction if machine state register
(problem state) bit, msr(pr) = 1. Allows software
priority change independent of msr(pr). If bit 19
is one, or 1,1,1 instruction sets low priority.
If bit 19 is zero, priority is set to low only if msr(pr) = 0
when the or 1,1,1 instruction is executed. See changing
priority with software, to be discussed later.
Thread switch priority enable
Forward progress count
Reserved in 64 bit register implementation
*A double-X load/store refers to loading or storing an elementary halfword, a word, or a double word, that crosses a doubleword boundary. A double-X load/store in this context is not a load or store of multiple words or a string of words.
Thread Switch Time-out Register
As discussed above, coarse grained multithreaded processors rely on long latency events to trigger thread switching. Sometimes during execution, a processor in a multiprocessor environment or a background thread in a multithreaded architecture, has ownership of a resource that can have only a single owner and another processor or active thread requires access to the resource before it can make forward progress. Examples include updating a memory page table or obtaining a task from a task dispatcher. The inability of the active thread to obtain ownership of the resource does not result in a thread switch event, nonetheless, the thread is spinning in a loop unable to do useful in work. In this case, the background thread that holds the resource does not obtain access to the processor so that it can free up the resource because it never encountered a thread switch event and does not become the active thread.
Allocating processing cycles among the threads is another concern; if software code running on a thread seldom encounters long latency switch events compared to software code running on the other threads in the same processor, that thread will get more than it's fair share of processing cycles. Yet another excessive delay that may exceed the maximum acceptable time is the latency of an inactive thread waiting to service an external interrupt within a limited period of time or some other event external to the processor. Thus, it becomes preferable to force a thread switch to the dormant thread after some time if no useful processing is being accomplished to prevent the system from hanging.
The logic to force a thread switch after a period of time is a thread switch time-out register 430 (FIG. 4), a decrementer, and a decrementer register to hold the decremented value. The thread switch time-out register 430 holds a thread switch time-out value. The thread switch time-out register 430 implementation used in this embodiment is shown in the following table:
Thread Switch Time-out Register Bits
Thread switch time-out value
The embodiment of the invention described herein does not implement a separate thread switch time-out register 430 for each thread, although that could be done to provide more flexibility. Similarly, if there are multiple threads, each thread need not have the same thread switch time-out value. Each time a thread switch occurs, the thread switch time-out value from the thread switch time-out register 430 is loaded by hardware into the decrement register. The decrement register is decremented once each cycle until the decrement register value equals zero, then a signal is sent to the thread switch controller 450 which forces a thread switch unless no other thread is ready to process instructions. For example, if all other threads in the system are waiting on a cache miss and are not ready to execute instructions, the thread switch controller 450 does not force a thread switch. If no other thread is ready to process instructions when the value in the decrement register reaches zero, the decremented value is frozen at zero until another thread is ready to process instructions, at which point a thread switch occurs and the decrement register is reloaded with a thread switch time-out value for that thread. Similarly, the decrement register could just as easily be named an increment register and when a thread is executing the register could increment up to some predetermined value when a thread switch would be forced.
The thread switch time-out register 430 can be written by a service processor as described above or by the processor itself with software code. The thread switch time-out value loaded into the thread switch time-out register 430 can be customized according to specific hardware configuration and/or specific software code to minimize wasted cycles resulting from unnecessary thread switching. Too high of a value in the thread switch time-out register 430 can result in reduced performance when the active thread is waiting for a resource held by another thread or if response latency for an external interrupt or some other event external to the processor is too long. Too high of a value can also prevent fairness if one thread experiences a high number of thread,switch events and the other does not. A thread switch time-out value twice to several times longer than the most frequent longest latency event that causes a thread switch is recommended, e.g., access to main memory. Forcing a thread switch after waiting the number of cycles specified in the thread switch time-out register 430 prevents system hangs due to shared resource contention, enforces fairness of processor cycle allocation between threads, and limits the maximum response latency to external interrupts and other events external to the processor.
Forward Progress Guarantee
That at least one instruction must be executed each time a thread switch occurs and a new thread becomes active is too restrictive in certain circumstances, such as when a single instruction generates multiple cache accesses and/or multiple cache misses. For example, a fetch instruction may cause an L1 I-cache 150 miss if the instruction requested is not in the cache; but when the instruction returns, required data may not be available in the L1 D-cache 120. Likewise, a miss in translation lookaside buffer 250 can also result in a data cache miss. So, if forward progress is strictly enforced, misses on subsequent accesses do not result in thread switches. A second problem is that some cache misses may require a large number of cycles to complete, during which time another thread may experience a cache miss at the same cache level which can be completed in much less time. If, when returning to the first thread, the strict forward progress is enforced, the processor is unable to switch to the thread with the shorter cache miss.
To remedy the problem of thrashing wherein each thread is locked in a repetitive cycle of switching threads without any instructions executing, there exists a forward progress count register 420 (FIG. 4) which allows up to a programmable maximum number of thread switches called the forward progress threshold value. After that maximum number of thread switches, an instruction must be completed before switching can occur again. In this way, thrashing is prevented. Forward progress count register 420 may actually be bits 30:31 in the thread switch control register 410 or a software programmable forward progress threshold register for the processor. The forward progress count logic uses bits 15:17 of the thread state registers 442, 444 that indicate the state of the threads and are allocated for the number of thread switches a thread has experienced without an instruction executing. Preferably, then these bits comprise the forward progress counter.
When a thread changes state invoking the thread switch algorithm, if at least one instruction has completed in the active thread, the forward-progress counter for the active thread is reset and the thread switch algorithm continues to compare thread states between the threads in the processor. If no instruction has completed, the forward-progress counter value in the thread state register of the active thread is compared to the forward progress threshold value. If the counter value is not equal to the threshold value, the thread switch algorithm continues to evaluate the thread states of the threads in the processor. Then if a thread switch occurs, the forward-progress counter is incremented. If, however, the counter value is equal to the threshold value, no thread switch will occur until an instruction can execute, i.e., until forward progress occurs. Note that if the threshold register has value zero, at least one instruction must complete within the active thread before switching to another thread. If each thread switch requires three processor cycles and if there are two threads and if the thread switch logic is programmed to stop trying to switch threads after five tries; then the maximum number of cycles that the processor will thrash is thirty cycles. One of skill in the art can appreciate that there a potential conflict exists between prohibiting a thread switch because no forward progress will be made on one hand and, on the other hand, forcing a thread switch because the time-out count has been exceeded. Such a conflict can easily be resolved according to architecture and software.
FIG. 6 is a flowchart of the forward progress count feature of thread switch logic 400 which prevents thrashing. At block 610, bits 15:17 in thread state register 442 pertaining to thread T0 are reset to state 111. Execution of this thread is attempted in block 620 and the state changes to 000. If an instruction successfully executes on thread T0, the state of thread T0 returns to 111 and remains so. If, however, thread T0 cannot execute an instruction, a thread switch occurs to thread T1, or another background thread if more than two threads are permitted in the processor architecture. When a thread switch occurs away from T1 or the other background thread and execution returns to thread T0, a second attempt to execute thread T0 occurs and the state of thread T0 becomes 001 as in block 630. Again, if thread T0 encounters a thread switch event, control of the processor is switched away from thread T0 to another thread. Similarly, whenever a thread switch occurs from the other thread, e.g., T1, back to thread T0, the state of T0 changes to 010 on this third attempt to execute T0 (block 640); to 011 on the fourth attempt to execute T0 (block 650), and to state 100 on the fifth attempt to execute T0 (block 660).
In this implementation, there are five attempts to switch to thread T0. After the fifth attempt or whenever the value of bits 15:17 in the thread state register (TSR) 442 is equal to the value of bits 30:31 plus one in the thread switch control register (TSC) 410, i.e., whenever TSC(30:31)+1=TSR (15:17), no thread switch away from thread T0 occurs. It will be appreciated that five attempts is an arbitrary number; the maximum number of allowable switches with unsuccessful execution, i.e., the forward progress threshold value, is programmable and it may be realized in certain architectures that five is too many switches, and in other architectures, five is too few. In any event, the relationship between the number of times that an attempt to switch to a thread with no instructions executing must be compared with a threshold value and once that threshold value has been reached, no thread switch occurs away from that thread and the processor waits until the latency associated with that thread is resolved. In the embodiment described herein, the state of the thread represented by bits 15:17 of the thread state register 442 is compared with bits 30:31 in the thread switch control register 410. Special handling for particular events that have extremely long latency, such as interaction with input/output devices, to prevent prematurely blocking thread switching with forward progress logic improves processor performance. One way to handle these extremely long latency events is to block the incrementing of the forward progress counter or ignore the output signal of the comparison between the forward progress counter and the threshold value if data has not returned. Another way to handle extremely long latency events is to use a separate larger forward progress count for these particular events.
Thread Switch Manager
The thread state for all software threads dispatched to the processor is preferably maintained in the thread state registers 442 and 444 of FIG. 4 as described. In a single processor one thread executes its instructions at a time and all other threads are dormant. Execution is switched from the active thread to a dormant thread when the active thread encounters a long-latency event as discussed above with respect to the forward progress register 420, the thread switch control register 410, or the thread switch time-out register 430. Independent of which thread is active, these hardware registers use conditions that do not dynamically change during the course of execution.
Flexibility to change thread switch conditions by a thread switch manager improves overall system performance. A software thread switch manager can alter the frequency of thread switching, increase execution cycles available for a critical task, and decrease the overall cycles lost because of thread switch latency. The thread switch manager can be programmed either at compile time or during execution by the operating system, e.g., a locking loop can change the frequency of thread switches; or an operating system task can be dispatched because a dormant thread in a lower priority state is waiting for an external interrupt or is otherwise ready. It may be advantageous to disallow or decrease the frequency of thread switches away from an active thread so that performance of the current instruction stream does not suffer the latencies resulting from switching into and out of it. Alternatively, a thread can forgo some or all of its execution cycles by essentially lowering its priority, and as a result, decrease the frequency of switches into it or increase the frequency of switches out of the thread to enhance overall system performance. The thread switch manager may also unconditionally force or inhibit a thread switch, or influence which thread is next selected for execution.
A multiple-priority thread switching scheme assigns a priority value to each thread to qualify the conditions that cause a switch. It may also be desirable in some cases to have the hardware alter thread priority. For instance, a low-priority thread may be waiting on some event, which when it occurs, the hardware can raise the priority of the thread to influence the response time of the thread to the event. Relative priorities between threads or the priority of a certain thread will influence the handling of such an event. The priorities of the threads can be adjusted by the thread switch manager software through the use of one or more instructions, or by hardware in response to an event. The thread switch manager alters the actions performed by the hardware thread switch logic to effectively change the relative priority of the threads.
Three priorities are used with the embodiment described herein of two threads and provides sufficient distinction between threads to allow tuning of performance without adversely affecting system performance. With three priorities, two threads can have an equal status of medium priority. The choice of three priorities for two threads is not intended to be limiting. In some architectures a “normal” state may be that one thread always has a higher priority than the other threads. It is intended to be within the scope of the invention to cover more than two threads of execution having one or multiple priorities that can be set in hardware or programmed by software.
The three priorities of each thread are high, medium, and low. When the priority of thread T0 is the same as thread T1, there is no effect on the thread switching logic. Both threads have equal priority so neither is given an execution time advantage. When the priority of thread T0 is greater than the priority of thread T1, thread switching from T0 to T1 is disabled for all L1 cache misses, i.e., data load, data store, and instruction fetch, because L1 cache misses are resolved much faster than other conditions such as L2 misses and translates. Thread T0 is given a better chance of receiving more execution cycles than thread T1 which allows thread T0 to continue execution so long as it does not waste an excessive number of execution cycles. The processor, however, will still relinquish control to thread T1 if thread T0 experiences a relatively long execution latency. Thread switching from T1 to T0 is unaffected, except that a switch occurs when dormant thread T0 is ready in which case thread T0 preempts thread Ti. This case would be expected to occur when thread T0 switches away because of an L2 cache miss or translation request, and the condition is resolved in the background while thread T0 is executing. The case of thread T0 having a priority less than thread T1 is analogous to the case above, with the thread designation reversed.
There are different possible approaches to implementing management of thread switching by changing thread priority. New instructions can be added to the processor architecture. Existing processor instructions having side effects that have the desired actions can also be used. Several factors influence the choice among the methods of allowing software control: (a) the ease of redefining architecture to include new instructions and the effect of architecture changes on existing processors; (b) the desirability of running identical software on different versions of processors; (c) the performance tradeoffs between using new, special purpose instructions versus reusing existing instructions and defining resultant side effects; (d) the desired level of control by the software, e.g., whether the effect can be caused by every execution of some existing instruction, such as a specific load or store, or whether more control is needed, by adding an instruction to the stream to specifically cause the effect.
The architecture described herein preferably takes advantage of an unused instruction whose values do not change the architected general purpose registers of the processor; this feature is critical for retrofitting multithreading capabilities into a processor architecture. Otherwise special instructions can be coded. The instruction is a “preferred nop” or 0,0,0; other instructions, however, can effectively act as a nop. By using different versions of the or instruction, or 0,0,0 or 1,1,1 etc. to alter thread priority, the same instruction stream may execute on a processor without adverse effects such as illegal instruction interrupts. An extension uses the state of the machine state register to alter the meaning of these instructions. For example, it may be undesirable to allow a user to code some or all of these thread priority instructions and access the functions they provide. The special functions they provide may be defined to occur only in certain modes of execution, they will have no effect in other modes and will be executed normally, as a nop.
One possible implementation, using a dual-thread multithreaded processor, uses three instructions which become part of the executing software itself to change the priority of itself:
tsop 1 or 1,1,1
- Switch to dormant thread
tsop 2 or 1,1,1
- Set active thread to LOW priority
- Switch to dormant thread
- NOTE: Only valid in privileged mode unless
tsop 3 or 2,2,2
- Set active thread to MEDIUM priority
tsop 4 or 3,3,3
- Set active thread to HIGH priority
- NOTE: Only valid in privileged mode
Instructions tsop 1 and tsop 2 can be the same instruction as embodied herein as or 1,1,1 but they can also be separate instructions. These instructions interact with bits 19 and 21 of the thread switch control register 410 and the problem/privilege bit of the machine state register as described herein. If bit 21 of the thread switch control register 410 has a value of one, the thread switch manager can set the priority of its thread to one of three priorities represented in the thread state register at bits 18:19. If bit 19 of the thread switch control register 410 has a value zero, then the instruction tsop 2 thread switch and thread priority setting is controlled by the problem/privilege bit of the machine state register. On the other hand, if bit 19 of the thread switch control register 410 has a value one, or if the problem/privilege bit of the machine state register has a value zero and the instruction or 1,1,1 is present in the code, the priority for the active thread is set to low and execution is immediately switched to the dormant or background thread if the dormant thread is enabled. The instruction or 2,2,2 sets the priority of the active thread to medium regardless of the value of the problem/privilege bit of the machine state register. And the instruction or 3,3,3, when the problem/privilege bit of the machine state register bit has a value of zero, sets the priority of the active thread to high. If bit 21 of the thread switch control register 320 is zero, the priority for both threads is set to medium and the effect of the or x,x,x instructions on the priority is blocked. If an external interrupt request is active, and if the corresponding thread's priority is low, that thread's priority is set to medium.
The events altered by the thread priorities are: (1) switch on L1 D-cache miss to load data; (2) switch on L1 D-cache miss for storing data; (3) switch on L1 I-cache miss on an instruction fetch; and (4) switch if the dormant thread in ready state. In addition, external interrupt activation may alter the corresponding thread's priority. The following table shows the effect of priority on conditions that cause a thread switch. A simple TSC entry in columns three and four means to use the conditions set forth in the thread switch control (TSC) register 410 to initiate a thread switch. An entry of TSC[0:2] treated as 0 means that bits 0:2 of the thread switch control register 410 are treated as if the value of those bits are zero for that thread and the other bits in the thread switch control register 410 are used as is for defining the conditions that cause thread switches. The phrase when thread T0 ready in column four means that a switch to thread T0 occurs as soon as thread T0 is no longer waiting on the miss event that caused it to be switched out. The phrase when thread T1 ready in column 3 means that a switch to thread T1 occurs as soon as thread T1 is no longer waiting on the miss event that caused it to be switched out. If the miss event is a thread switch time-out, there is no guarantee that the lower priority thread completes an instruction before the higher priority thread switches back in.
TSC[0:2] treated as 0
TSC or if T0 ready
TSC[0:2] treated as 0
TSC or if T0 ready
TSC or if T1 ready
TSC[0:2] treated as 0
TSC[0:2] treated as 0
TSC or if T0 ready
TSC or if T1 ready
TSC[0:2] treated as 0
TSC or if T1 ready
TSC[0:2] treated as 0
It is recommended that a thread doing no productive work be given low priority to avoid a loss in performance even if every instruction in the idle loop causes a thread switch. Yet, it is still important to allow hardware to alter thread priority if an external interrupt is requested to a thread set at low priority. In this case the thread is raised to medium priority, to allow a quicker response to the interrupt. This allows a thread waiting on an external event to set itself at low priority, where it will stay until the event is signalled.
While the invention has been described in connection with what is presently considered the most practical and preferred embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3373408||16 Apr 1965||12 Mar 1968||Rca Corp||Computer capable of switching between programs without storage and retrieval of the contents of operation registers|
|US3566357||5 Jul 1966||23 Feb 1971||Rca Corp||Multi-processor multi-programed computer system|
|US3568173||19 Aug 1968||2 Mar 1971||Scient Data Systems Inc||Memory stroage element drive circuit|
|US3594732||29 Oct 1969||20 Jul 1971||Scient Data Systems Inc||General purpose digital computer|
|US3728692||31 Aug 1971||17 Apr 1973||Ibm||Instruction selection in a two-program counter instruction unit|
|US3771138||31 Aug 1971||6 Nov 1973||Ibm||Apparatus and method for serializing instructions from two independent instruction streams|
|US3916383||20 Feb 1973||28 Oct 1975||Memorex Corp||Multi-processor data processing system|
|US3980991||29 Nov 1974||14 Sep 1976||Ing. C. Olivetti & C., S.P.A.||Apparatus for controlling microprogram jumps in a microprogrammable electronic computer|
|US3980992||26 Nov 1974||14 Sep 1976||Burroughs Corporation||Multi-microprocessing unit on a single semiconductor chip|
|US4047161 *||30 Apr 1976||6 Sep 1977||International Business Machines Corporation||Task management apparatus|
|US4084228||2 Dec 1974||11 Apr 1978||Compagnie Honeywell Bull||Process management structures and hardware/firmware control|
|US4229790||16 Oct 1978||21 Oct 1980||Denelcor, Inc.||Concurrent task and instruction processor and method|
|US4320453||2 Nov 1978||16 Mar 1982||Digital House, Ltd.||Dual sequencer microprocessor|
|US4384324||6 May 1980||17 May 1983||Burroughs Corporation||Microprogrammed digital data processing system employing tasking at a microinstruction level|
|US4493020||29 Jul 1980||8 Jan 1985||Burroughs Corporation||Microprogrammed digital data processor employing microinstruction tasking and dynamic register allocation|
|US4532587||26 Aug 1981||30 Jul 1985||Texas Instruments Incorporated||Single chip processor connected to an external memory chip|
|US4590555||11 Dec 1980||20 May 1986||Compagnie Internationale Pour L'informatique Cii-Honeywell Bull (Societe Anonyme)||Apparatus for synchronizing and allocating processes among several processors of a data processing system|
|US4692861||11 Apr 1984||8 Sep 1987||Inmos Limited||Microcomputer with interprocess communication|
|US4704678||16 Nov 1983||3 Nov 1987||Inmos Limited||Function set for a microcomputer|
|US4829425||1 Mar 1988||9 May 1989||Intel Corporation||Memory-based interagent communication mechanism|
|US4853849||17 Dec 1986||1 Aug 1989||Intel Corporation||Multi-tasking register set mapping system which changes a register set pointer block bit during access instruction|
|US4939755||12 Nov 1987||3 Jul 1990||Nec Corporation||Timer/counter using a register block|
|US5050070||29 Feb 1988||17 Sep 1991||Convex Computer Corporation||Multi-processor computer system having self-allocating processors|
|US5079725||17 Nov 1989||7 Jan 1992||Ibm Corporation||Chip identification method for use with scan design systems and scan testing techniques|
|US5103394||21 Dec 1989||7 Apr 1992||Hewlett-Packard Company||Software performance analyzer|
|US5148536||25 Jul 1988||15 Sep 1992||Digital Equipment Corporation||Pipeline having an integral cache which processes cache misses and loads data in parallel|
|US5159686||7 Mar 1991||27 Oct 1992||Convex Computer Corporation||Multi-processor computer system having process-independent communication register addressing|
|US5179702||11 Jun 1990||12 Jan 1993||Supercomputer Systems Limited Partnership||System and method for controlling a highly parallel multiprocessor using an anarchy based scheduler for parallel execution thread scheduling|
|US5197138||26 Dec 1989||23 Mar 1993||Digital Equipment Corporation||Reporting delayed coprocessor exceptions to code threads having caused the exceptions by saving and restoring exception state during code thread switching|
|US5287508||7 Apr 1992||15 Feb 1994||Sun Microsystems, Inc.||Method and apparatus for efficient scheduling in a multiprocessor system|
|US5339415||9 Nov 1992||16 Aug 1994||Cray Research, Inc.||Dual level scheduling of processes to multiple parallel regions of a multi-threaded program on a tightly coupled multiprocessor computer system|
|US5353418||21 May 1990||4 Oct 1994||Massachusetts Institute Of Technology||System storing thread descriptor identifying one of plural threads of computation in storage only when all data for operating on thread is ready and independently of resultant imperative processing of thread|
|US5357617||22 Nov 1991||18 Oct 1994||International Business Machines Corporation||Method and apparatus for substantially concurrent multiple instruction thread processing by a single pipeline processor|
|US5361337||8 May 1992||1 Nov 1994||Sun Microsystems, Inc.||Method and apparatus for rapidly switching processes in a computer system|
|US5404469||25 Feb 1992||4 Apr 1995||Industrial Technology Research Institute||Multi-threaded microprocessor architecture utilizing static interleaving|
|US5408671||19 Mar 1992||18 Apr 1995||Nec Corporation||System for controlling shared registers|
|US5423008||3 Aug 1992||6 Jun 1995||Silicon Graphics, Inc.||Apparatus and method for detecting the activities of a plurality of processors on a shared bus|
|US5430850||22 Jul 1991||4 Jul 1995||Massachusetts Institute Of Technology||Data processing system with synchronization coprocessor for multiple threads|
|US5471593||21 Jan 1994||28 Nov 1995||Branigin; Michael H.||Computer processor with an efficient means of executing many instructions simultaneously|
|US5490272||28 Jan 1994||6 Feb 1996||International Business Machines Corporation||Method and apparatus for creating multithreaded time slices in a multitasking operating system|
|US5499349||3 Oct 1994||12 Mar 1996||Massachusetts Institute Of Technology||Pipelined processor with fork, join, and start instructions using tokens to indicate the next instruction for each of multiple threads of execution|
|US5515538||29 Mar 1994||7 May 1996||Sun Microsystems, Inc.||Apparatus and method for interrupt handling in a multi-threaded operating system kernel|
|US5524250||23 Dec 1994||4 Jun 1996||Silicon Graphics, Inc.||Central processing unit for processing a plurality of threads using dedicated general purpose registers and masque register for providing access to the registers|
|US5530866||13 Oct 1994||25 Jun 1996||Tera Computer Company||Register allocation methods having upward pass for determining and propagating variable usage information and downward pass for binding; both passes utilizing interference graphs via coloring|
|US5535361||24 May 1993||9 Jul 1996||Matsushita Electric Industrial Co., Ltd.||Cache block replacement scheme based on directory control bit set/reset and hit/miss basis in a multiheading multiprocessor environment|
|US5553305||15 Jul 1992||3 Sep 1996||International Business Machines Corporation||System for synchronizing execution by a processing element of threads within a process using a state indicator|
|US5574939||29 Jun 1995||12 Nov 1996||Massachusetts Institute Of Technology||Multiprocessor coupling system with integrated compile and run time scheduling for parallelism|
|US5613114||15 Apr 1994||18 Mar 1997||Apple Computer, Inc||System and method for custom context switching|
|US5630128||9 Aug 1991||13 May 1997||International Business Machines Corporation||Controlled scheduling of program threads in a multitasking operating system|
|US5630130 *||10 Dec 1993||13 May 1997||Centre Electronique Horloger S.A.||Multi-tasking low-power controller having multiple program counters|
|US5630136||9 Jun 1995||13 May 1997||Sun Microsystems, Inc.||Method and apparatus for serializing access to multithreading unsafe resources|
|US5809450 *||26 Nov 1997||15 Sep 1998||Digital Equipment Corporation||Method for estimating statistics of properties of instructions processed by a processor pipeline|
|US5826081||6 May 1996||20 Oct 1998||Sun Microsystems, Inc.||Real time thread dispatcher for multiprocessor applications|
|US5864701 *||14 Feb 1997||26 Jan 1999||Integrated Device Technology, Inc.||Apparatus and method for managing interrupt delay associated with mask flag transition|
|US5867644 *||10 Sep 1996||2 Feb 1999||Hewlett Packard Company||System and method for on-chip debug support and performance monitoring in a microprocessor|
|US5933627||1 Jul 1996||3 Aug 1999||Sun Microsystems||Thread switch on blocked load or store using instruction thread field|
|US5996058 *||19 Aug 1996||30 Nov 1999||Samsung Electronics Company, Ltd.||System and method for handling software interrupts with argument passing|
|US6016542 *||31 Dec 1997||18 Jan 2000||Intel Corporation||Detecting long latency pipeline stalls for thread switching|
|US6049867 *||4 Aug 1997||11 Apr 2000||International Business Machines Corporation||Method and system for multi-thread switching only when a cache miss occurs at a second or higher level|
|US6088788 *||27 Dec 1996||11 Jul 2000||International Business Machines Corporation||Background completion of instruction and associated fetch request in a multithread processor|
|US6295600||28 Jun 1999||25 Sep 2001||Sun Microsystems, Inc.||Thread switch on blocked load or store using instruction thread field|
|US6341347||11 May 1999||22 Jan 2002||Sun Microsystems, Inc.||Thread switch logic in a multiple-thread processor|
|EP0474297A2||2 Sep 1991||11 Mar 1992||Philips Electronics N.V.||Very long instruction word machine for efficient execution of programs with conditional branches|
|EP0479390A2||3 Oct 1991||8 Apr 1992||Philips Electronics N.V.||Processing device including a memory circuit and a group of functional units|
|EP0527392A2||29 Jul 1992||17 Feb 1993||International Business Machines Corporation||Preemptive and non pre-emptive scheduling and executing of program threads in a multitasking operating system|
|EP0605927A1||24 Dec 1993||13 Jul 1994||Philips Electronics N.V.||Improved very long instruction word processor architecture|
|EP0617361A2||18 Mar 1994||28 Sep 1994||Digital Equipment Corporation||Scheduling method and apparatus for a communication network|
|EP0747816A2||31 May 1996||11 Dec 1996||International Business Machines Corporation||Method and system for high performance multithread operation in a data processing system|
|GB2234613A||Title not available|
|1||"Honeywell 800 System", from the Collections of the University of Toledo Libraries, PP. 0-42.|
|2||"SDS Sigma 7 Computer Reference Manual", Scientific Data Systems, May 1966, PP. 0-95.|
|3||"Sigma 7 in the News", Scientific Data Systems, PP. 0-27.|
|4||"Sigma 7", Scientific Data Systems, PP. 0-21.|
|5||"Sigma 7", Scientific Data Systems, PP. 0-24.|
|6||*||Agarwal et al., "Sparcle: An Envolutionary Processor Design for Large-Scale Multiprocessors," IEEE, pp. 48-61, Jun. 1993.*|
|7||Agarwal, A., "Analysis of Cache Performance for Operating Systems and Multiprogramming", Dissertation Abstracts International, vol. 48, No. 7, Jan. 1988, PP., 2024-B.|
|8||Anderson, D. W. et al, The IBM System/360 Model 91: Machine Philosophy and Instruction-Handling IBM Journal of Research and Development, vol. 11, 1967, PP. 8-24.|
|9||Ararwal, A. et al., "APRIL: A Processor Architecture for Multiprocessing", Proceedings of the 17th Ar Symposium on Computer Architecture, May 1990, PP. 104-114.|
|10||Arvind et al, "A Critique of multiprocessing von Neumann Style", Proceedings of the 10th ACM International Symposium on computer Architecture, Jun. 1983, PP. 426-436.|
|11||Aschenbrenner, R. et al., "Intrinsic Multiprocessing", AFIPS Conference Proceedings, vol. 30, 1967 Spring Joint Computer Conference, PP. 81-86.|
|12||Athas, William C. et al, "Multicomputers: Message-Passing Concurrent Computers", Computer, vol. 21, No. 8, Aug. 1988, PP. 9-24.|
|13||Atwood, J. William; Computer Magazine, Oct. 1976, pp. 18-26, "Concurrency in Operating Systems".|
|14||Bernard Karl Gunther, "Superscalar Performance in a Multithreaded Microprocessor," University of Tasmania Hobart, Chapters 2 and 3, Dec. 1993.|
|15||Bolychevsky, A. et al; IEE Proc-Computer Digit Tech, vol. 143, No. 5, Sep. 1996, pp. 309-317, "Dynamic Scheduling in RISC Architectures".|
|16||Bradley J. Kish and Bruno R. Preiss, "Hobbes: A Multi-Threaded Supercalar Architecture," University of Waterloo, 1994.|
|17||Buehrer, Richard et al, "Incorporating Data Flow Ideas into von Neuman Processors for Parallel Execution", IEEE Transactions on Computers, vol. C-36, No. 12, Dec. 1987, PP. 1515-1522.|
|18||Chang, Yen et al, "Use Structured Arrays for High-Performance Data Processing", Electtonic Design News, vol. 32, No. 8, Apr. 1987, PP. 177-184.|
|19||Cheng, Ching Yuang et al, "RPM: A Fast RISC Type Prolog Machine", Proceedings VLSI and Computers, First Int'l. Conference onComputer Technology, Systems and Applications, Hamburg, May 11-15, 1987.|
|20||Cook, Robert W. et al, "System Design of a Dynamic Microprocessor", IEEE Transactions on Compute: vol. C-19, No. 3, MAr. 1970, PP. 213-222.|
|21||*||Cui et al., "Parallel Replacement Mechanism for MultiThread," IEEE, pp. 338-344, Mar. 1997.*|
|22||Cull, Thomas C., The Honeywell 8200, Datamation, vol. 11, No. 7, Jul. 1965, PP. 90-91.|
|23||Culler, Arvind and David E., "Dataflow Architectures", Annual Review Computer Science, 1986, pp. 225-253.|
|24||Dally, William J. et al, "Architecture of a Message-Driven Processor", 14th Annual Interantional Symposium on Computer Architecture, Jun. 1987, PP. 189-196.|
|25||Dreyfus, P., "Programming Design Features of the GAMMA 60 Computer", Proceedings of the Eastern Joint Computer Conference, Dec. 3-5, 1958, PP. 174-181.|
|26||Dubey, Pradeep K. et al; IEEE, 1995, pp. 110-122, "Analytic Performance Modeling for a Spectrum of Multithreaded Processor Architectures".|
|27||Eickememyer et al, "Evaluation of Multithreaded Uniprocessors For Commerical Application Environments", 23RD International Symposium On Computer Architecture, PP. 203-212, May 1996.|
|28||Elkateeb, Ali et al; IEEE Pacific Rim Conference 1993, pp. 141-144, "A Task Allocation by Priority Strategy for RISC Architecture Supported with Non-Overlapped Multiple Register Set: A Complexity Study ".|
|29||Fiske, Stuart et al; Future Generation Computer Systems II, Oct. 1995, No. 6, pp. 503-518, "Thread Prioritization: A Thread Scheduling Mechanism for Multiple-Contett Parallel Processors".|
|30||Flynn, Michael J. et al, An Unconventional Computer Architecture: Shared Resource Multiprocessing Computer, Mar./Apr. 1972, PP. 20-28.|
|31||Foster, C. C., "Uncoupling Central Processor and Storage Device Speeds", Teh Computer Journal, vol. 14, No. 1, PP. 45-48.|
|32||Fujita, Tetsuya "A Multithreaded Processor Architecture for Parallel Symbolic Computation", MIT Laboratory for Computer Science, MIT/LCS/TM-338, Sep. 1987, PP. 0-71.|
|33||Ghose, Kanad et al, "Using Method Lookup Caches and Register Windowing to Speed Up Dynamically-Bound Object-Oriented Applications", Proceedings of the 22nd EUROMICRO Conference, EUROMICRO 96, PP. 441-448.|
|34||Halstead, Robert H. Jr. et al, "MASA: A Multithreaded Processor Architecture for Parallel Symbolic Computing", 15th Annual International Symposium on Computer Architecture, 1988, PP. 443-451.|
|35||Hath, T.F. Jr. et al, "Hardware/Software Interaction on the Honeywell Model 8200", AFIPS Confe: Proceedings, vol. 33, PArt 1, 1968, Fall Joint Computer Conference, PP. 891-901.|
|36||Hidaka, Yasuo et al, "Multiple Threads in Cyclic Register Windows", Proceedings of the 20th Annual Int'l. Symposium on Computer Architecture, May 193, PP. 131-142.|
|37||Holfmann, Rick "Dual-Port Register File with Overlapped Register Windowing", IBM Technical Report TR54.660, Apr. 20, 1992, PP. 0-37.|
|38||Howe's Free On-Line Dictionary of Computing (http://www.instantweb.com/foldoc/), background, foreground, Oct. 1994.|
|39||IBM Technical Disclosure Bulletin, vol. 33, No. 9, Feb. 1991, pp. 446-447, "Generic Unix Off-Level Interrupt Handling Invocation Algorithm".|
|40||IBM Technical Disclosure Bulletin, vol. 38, No. 5, May 1995, pp. 271-276, "Deterministic Priority Inversion Method for Personal Computers".|
|41||IBM Technical Disclosure Bulletin, vol. 38, No. 9, Sep. 1995, pp. 563-565, "Deadline Monotonic Server".|
|42||IBM Technical Disclosure Bulletin, vol. 39, No. 08, Aug. 1996, PP 113-116, Entitled "Algorithm For Instruction Cache Arbitration Among Multiple Instruction Streams".|
|43||Inohara, Shigekazu et al; IEEE, 1993, pp. 149-155, "Unstable Threads Kernel Interface for Minimizing the Overhead of Thread Switching".|
|44||Kaminski, W. J. et al, "Developing a Multiple-Instruction-Stream Single-Chip Processor", Computer, Dec. 1979, PP. 66-76.|
|45||Kishi, Masasuke et al, "DDDP: A Distributed Data Driven Processor", Proceedings of 10th International Symposium on Computer Architecture, 1983, PP. 236-242.|
|46||Konsek, Marian B. et al, "Context Switching With Multiple Register Windows: A RISC Performance Study", Stanford Technical Report No. UIUCDS-R-87-1377, Oct. 1987, PP. 0-32.|
|47||Kuehn, James T. et al, "The Horizon Supercomputing System: Architecture and Software", Proceedings of Supercomputing '88, PP. 28-34.|
|48||Lorin, Harold "Introduction to Computer Architecture and Organization", A Wiley-Interscience Publication, John Wiley & Sons, PP. 41-53.|
|49||Markoff, John "RISC Chips-RISC Means Longer Programs but Faster Execution", BYTE, vol. 9, No. 12, Dec. 1984, PP. 191-206.|
|50||Markoff, John "RISC Chips—RISC Means Longer Programs but Faster Execution", BYTE, vol. 9, No. 12, Dec. 1984, PP. 191-206.|
|51||Mayhew, David et al, "Overlapping Register Windows with Transparent Spill and In-Line Subroutine Expansion", Conference Proceedings IEEE Southeastcon '87, Apr. 1987, PP. 627-630.|
|52||Mike Johnson, "Superscalar Microprocessor Design," Prentice Hall, pp. 21, 46-47, 233-235, 268-272, 1991.|
|53||Miller, Edward F. Jr., "A Multiple-Stream Registerless Shared-Resource Processor", IEEE Transcations on Computers, vol. c-23, No. 3, Mar. 1974, PP. 277-285.|
|54||Nisenoff, N., "Scratchpad Memories at Honeywell: Past, Present, and Future", AFIPS Conference Proceedings, vol. 27, Part 1, 1965, Fall Joint Computer Conference, PP. 679-688.|
|55||Ohr, Stephen "RISC Machines", Electronic Design, vol. 33, Jan. 10, 1985, PP. 174-190.|
|56||Okamoto, Kazuaki et al, "Multithread Execution Mechanisms on RICA-1 for Massively Parallel Computation", Proceedings of the 1996 Conference on Parallel Architectures and Compilation Techniques (PAT '96), PP. 116-121.|
|57||Papadopoulos, Gregory M., "Implementation of a General Purpose Dataflow Multiprocessor", MIT Electrical Engineering and Computer Science PHD Thesis, Aug. 1988, PP. 0-155.|
|58||Ragan-Kelley, Robert et al, "Applying RISC Theory to a Large Computer", Computer Design, vol. 22, No. 13, Nov. 1983, PP. 191-198.|
|59||Seitz, Charles L., "The Cosmic Cube", Communications of the ACM, vol. 28, No. 1, Jan. 1985, PP. 22-33.|
|60||Shar, Leonard E. et al, "A Multiminiprocessor System Implemented Through Pipelining", Computer, Feb. 1974, PP. 42-51.|
|61||Smith, Bob "Chips in Transition", PC Tech Journal, vol. 4, No. 4, Apr. 1986, PP. 56-63.|
|62||Smith, Burton J. "A Pipelined, Shared Resource MIMD Computer", Proceedings of the 1978 Int'l. Conference on Parallel Processing, Aug. 1978, pp. 6-8.|
|63||Smith, Burton J. "Architecture and Applications of the HEP Multiprocessor Computer system", Proceedings of SPIE-The Int'l. Society for Optical Engineering, Vo. 298, Aug. 1981, PP. 241-248.|
|64||Smith, Burton J. "Architecture and Applications of the HEP Multiprocessor Computer system", Proceedings of SPIE—The Int'l. Society for Optical Engineering, Vo. 298, Aug. 1981, PP. 241-248.|
|65||Smith, Burton J. et al, Architecture and Applications of the HEP Multiprocessor Computer System, Proceedings of the Conference on Peripheral Array Processors, Oct. 1982, PP. 159-170.|
|66||Snelling, David F. et al, "MIMD Processing and the Denelcor HEP", Multiprocessing in Meteorological Models, Springer-Verlag, PP. 89-95.|
|67||Song P., "Multithreading Comes Of Age", Microdesign Resources, Microprocessor Report, PP. 13-18, Jul. 14, 1997.|
|68||Sporer, michael et al, "An Introduction to the Architecture of the Stellar Graphics Supercomputer" Proceedings of Compcon Spring '88, 33rd IEEE Computer Society Int'l. Confernece, PP. 464-467.|
|69||*||Thekkath et al., "The Effectiveness of Multiple Hardware Contexts," ACM 0-89791-660-3/94/0010 (ASPLOS VI), pp 328-337, Oct. 1994.*|
|70||Thistle, Mark R. et al, "Processor Architecture for Horizon", IEEE Supercomputing '88, PP. 35-41.|
|71||Thornton, J. E., "Design of a Computer-The Control Data 6600", Scott, Foresman and Company, 1970, PP. 115-175.|
|72||Thornton, J. E., "Design of a Computer—The Control Data 6600", Scott, Foresman and Company, 1970, PP. 115-175.|
|73||Tokuda, Hideyuki et al; IEEE, 1989, pp. 348-359, "Priority Inversions in Real-Time Communication".|
|74||Topham, Nigel P. et al, "context Flow: An Alternative to Conventional Pipelined Architectures", The Journal of Supercomputing, vol. 2, No. 1, 1988, PP. 29-53.|
|75||Wallich, Paul "Toward Simpler, Faster Computers", IEEE Spectrum, vol. 22, No. 8, Aug. 8, 1985, PP. 38-45.|
|76||Weinberg, William; Posix, Real-Time Magazine 97-2, pp. 51-54, "Meeting Real-Time Performance Goals With Kernel Threads".|
|77||Whitby-Stevens, Colin "The Transputer", IEEE Proceedings of 12th Annual International Symposium Computer Architecture, Jun. 1985, PP. 292-300.|
|78||Willis, I. C. et al, "RapidGraph: An Object Flow Multiprocessor", Proceedings of the International Workshop on Programming Environments for Intelligent Systems, 1985.|
|79||Wilson, Ron "Higher Speeds Push Embedded Systems to Multiprocessing", Computer Design, Jul. 1, 1989, PP. 72-83.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US6883107 *||8 Mar 2002||19 Apr 2005||Intel Corporation||Method and apparatus for disabling a clock signal within a multithreaded processor|
|US6892331 *||17 Jan 2002||10 May 2005||International Business Machines Corporation||Method and system for error detection in a managed application environment|
|US6944736 *||11 Jan 2002||13 Sep 2005||Hewlett-Packard Development Company, L.P.||Managing latencies in accessing memory of computer systems|
|US6971103 *||1 Apr 2003||29 Nov 2005||Sandbridge Technologies, Inc.||Inter-thread communications using shared interrupt register|
|US7013400 *||24 Apr 2003||14 Mar 2006||International Business Machines Corporation||Method for managing power in a simultaneous multithread processor by loading instructions into pipeline circuit during select times based on clock signal frequency and selected power mode|
|US7082601 *||17 Jul 2002||25 Jul 2006||Nec Corporation||Multi-thread execution method and parallel processor system|
|US7096470 *||19 Sep 2002||22 Aug 2006||International Business Machines Corporation||Method and apparatus for implementing thread replacement for optimal performance in a two-tiered multithreading structure|
|US7130986 *||30 Jun 2003||31 Oct 2006||Intel Corporation||Determining if a register is ready to exchange data with a processing element|
|US7178005||30 Jun 2004||13 Feb 2007||Sun Microsystems, Inc.||Efficient implementation of timers in a multithreaded processor|
|US7181741 *||30 Jan 2003||20 Feb 2007||Hewlett-Packard Development Company, L.P.||Apparatus and method to minimize blocking overhead in upcall based MxN threads|
|US7185178||30 Jun 2004||27 Feb 2007||Sun Microsystems, Inc.||Fetch speculation in a multithreaded processor|
|US7216216||30 Jun 2004||8 May 2007||Sun Microsystems, Inc.||Register window management using first pipeline to change current window and second pipeline to read operand from old window and write operand to new window|
|US7216223||30 Apr 2004||8 May 2007||Hewlett-Packard Development Company, L.P.||Configuring multi-thread status|
|US7287146 *||2 Feb 2005||23 Oct 2007||Nec Corporation||Array-type computer processor|
|US7321965 *||27 Aug 2004||22 Jan 2008||Mips Technologies, Inc.||Integrated mechanism for suspension and deallocation of computational threads of execution in a processor|
|US7330988||30 Jun 2004||12 Feb 2008||Sun Microsystems, Inc.||Method and apparatus for power throttling in a multi-thread processor|
|US7343474||30 Jun 2004||11 Mar 2008||Sun Microsystems, Inc.||Minimal address state in a fine grain multithreaded processor|
|US7353364||30 Jun 2004||1 Apr 2008||Sun Microsystems, Inc.||Apparatus and method for sharing a functional unit execution resource among a plurality of functional units|
|US7370243||30 Jun 2004||6 May 2008||Sun Microsystems, Inc.||Precise error handling in a fine grain multithreaded multicore processor|
|US7370325 *||19 Feb 1999||6 May 2008||Lucent Technologies Inc.||Eager evaluation of tasks in a workflow system|
|US7373489||30 Jun 2004||13 May 2008||Sun Microsystems, Inc.||Apparatus and method for floating-point exception prediction and recovery|
|US7376954||10 Oct 2003||20 May 2008||Mips Technologies, Inc.||Mechanisms for assuring quality of service for programs executing on a multithreaded processor|
|US7383403||30 Jun 2004||3 Jun 2008||Sun Microsystems, Inc.||Concurrent bypass to instruction buffers in a fine grain multithreaded processor|
|US7401206||30 Jun 2004||15 Jul 2008||Sun Microsystems, Inc.||Apparatus and method for fine-grained multithreading in a multipipelined processor core|
|US7415556 *||23 Oct 2003||19 Aug 2008||International Business Machines Corporation||Exclusion control|
|US7418585 *||11 Jan 2006||26 Aug 2008||Mips Technologies, Inc.||Symmetric multiprocessor operating system for execution on non-independent lightweight thread contexts|
|US7424599||27 Aug 2004||9 Sep 2008||Mips Technologies, Inc.||Apparatus, method, and instruction for software management of multiple computational contexts in a multithreaded microprocessor|
|US7426630||30 Jun 2004||16 Sep 2008||Sun Microsystems, Inc.||Arbitration of window swap operations|
|US7426732||10 Oct 2003||16 Sep 2008||Cray Inc.||Placing a task of a multithreaded environment in a known state|
|US7434000||9 Dec 2004||7 Oct 2008||Sun Microsystems, Inc.||Handling duplicate cache misses in a multithreaded/multi-core processor|
|US7437538||30 Jun 2004||14 Oct 2008||Sun Microsystems, Inc.||Apparatus and method for reducing execution latency of floating point operations having special case operands|
|US7478225||30 Jun 2004||13 Jan 2009||Sun Microsystems, Inc.||Apparatus and method to support pipelining of differing-latency instructions in a multithreaded processor|
|US7478394 *||4 Jun 2001||13 Jan 2009||Hewlett-Packard Development Company, L.P.||Context-corrupting context switching|
|US7493621||18 Dec 2003||17 Feb 2009||International Business Machines Corporation||Context switch data prefetching in multithreaded computer|
|US7523330||30 Jun 2004||21 Apr 2009||Sun Microsystems, Inc.||Thread-based clock enabling in a multi-threaded processor|
|US7523455 *||3 May 2002||21 Apr 2009||Hewlett-Packard Development Company, L.P.||Method and system for application managed context switching|
|US7533248 *||30 Jun 2004||12 May 2009||Sun Microsystems, Inc.||Multithreaded processor including a functional unit shared between multiple requestors and arbitration therefor|
|US7536690 *||16 Sep 2003||19 May 2009||Cray Inc.||Deferred task swapping in a multithreaded environment|
|US7594089 *||30 Sep 2004||22 Sep 2009||Mips Technologies, Inc.||Smart memory based synchronization controller for a multi-threaded multiprocessor SoC|
|US7596683||11 Jul 2007||29 Sep 2009||Intel Corporation||Switching processor threads during long latencies|
|US7610473||27 Aug 2004||27 Oct 2009||Mips Technologies, Inc.||Apparatus, method, and instruction for initiation of concurrent instruction streams in a multithreading microprocessor|
|US7617380 *||25 Aug 2005||10 Nov 2009||Broadcom Corporation||System and method for synchronizing translation lookaside buffer access in a multithread processor|
|US7617499||18 Dec 2003||10 Nov 2009||International Business Machines Corporation||Context switch instruction prefetching in multithreaded computer|
|US7631125 *||30 Sep 2005||8 Dec 2009||Intel Corporation||Dynamically migrating channels|
|US7634774 *||13 Sep 2004||15 Dec 2009||Integrated Device Technology, Inc.||System and method of scheduling computing threads|
|US7650484 *||3 Feb 2005||19 Jan 2010||Nec Corporation||Array—type computer processor with reduced instruction storage|
|US7669203 *||19 Dec 2003||23 Feb 2010||Intel Corporation||Virtual multithreading translation mechanism including retrofit capability|
|US7676655||30 Jun 2004||9 Mar 2010||Sun Microsystems, Inc.||Single bit control of threads in a multithreaded multicore processor|
|US7676660||3 Dec 2007||9 Mar 2010||Mips Technologies, Inc.||System, method, and computer program product for conditionally suspending issuing instructions of a thread|
|US7676664 *||23 Dec 2006||9 Mar 2010||Mips Technologies, Inc.||Symmetric multiprocessor operating system for execution on non-independent lightweight thread contexts|
|US7681018||12 Jan 2001||16 Mar 2010||Intel Corporation||Method and apparatus for providing large register address space while maximizing cycletime performance for a multi-threaded register file set|
|US7694304||27 Aug 2004||6 Apr 2010||Mips Technologies, Inc.||Mechanisms for dynamic configuration of virtual processor resources|
|US7702887||30 Jun 2004||20 Apr 2010||Sun Microsystems, Inc.||Performance instrumentation in a fine grain multithreaded multicore processor|
|US7711931||30 Sep 2004||4 May 2010||Mips Technologies, Inc.||Synchronized storage providing multiple synchronization semantics|
|US7725689 *||23 Dec 2006||25 May 2010||Mips Technologies, Inc.||Symmetric multiprocessor operating system for execution on non-independent lightweight thread contexts|
|US7725697 *||23 Dec 2006||25 May 2010||Mips Technologies, Inc.|
|US7730291 *||23 Dec 2006||1 Jun 2010||Mips Technologies, Inc.|
|US7743235||6 Jun 2007||22 Jun 2010||Intel Corporation||Processor having a dedicated hash unit integrated within|
|US7747771||30 Jun 2004||29 Jun 2010||Oracle America, Inc.||Register access protocol in a multihreaded multi-core processor|
|US7774393||30 Jun 2004||10 Aug 2010||Oracle America, Inc.||Apparatus and method for integer to floating-point format conversion|
|US7793023||4 Apr 2008||7 Sep 2010||International Business Machines Corporation||Exclusion control|
|US7823158||18 Aug 2005||26 Oct 2010||International Business Machines Corporation||Adaptive scheduling and management of work processing in a target context in resource contention|
|US7827555||14 Jan 2005||2 Nov 2010||Integrated Device Technology, Inc.||Scheduler for a multiprocessing computing system|
|US7831980 *||25 Mar 2008||9 Nov 2010||International Business Machines Corporation||Scheduling threads in a multi-processor computer|
|US7836450||11 Jan 2006||16 Nov 2010||Mips Technologies, Inc.|
|US7849297||20 Dec 2005||7 Dec 2010||Mips Technologies, Inc.||Software emulation of directed exceptions in a multithreading processor|
|US7856633||24 Mar 2000||21 Dec 2010||Intel Corporation||LRU cache replacement for a partitioned set associative cache|
|US7861063||30 Jun 2004||28 Dec 2010||Oracle America, Inc.||Delay slot handling in a processor|
|US7870553||11 Jan 2006||11 Jan 2011||Mips Technologies, Inc.|
|US7873816||20 Nov 2008||18 Jan 2011||International Business Machines Corporation||Pre-loading context states by inactive hardware thread in advance of context switch|
|US7890734||30 Jun 2004||15 Feb 2011||Open Computing Trust I & II||Mechanism for selecting instructions for execution in a multithreaded processor|
|US7890735||23 Aug 2006||15 Feb 2011||Texas Instruments Incorporated||Multi-threading processors, integrated circuit devices, systems, and processes of operation and manufacture|
|US7917907 *||23 Mar 2005||29 Mar 2011||Qualcomm Incorporated||Method and system for variable thread allocation and switching in a multithreaded processor|
|US7921365 *||15 Feb 2005||5 Apr 2011||Microsoft Corporation||System and method for browsing tabbed-heterogeneous windows|
|US7926035||24 Apr 2007||12 Apr 2011||Microsoft Corporation||Testing multi-thread software using prioritized context switch limits|
|US7941642||30 Jun 2004||10 May 2011||Oracle America, Inc.||Method for selecting between divide instructions associated with respective threads in a multi-threaded processor|
|US7991983||3 Jun 2009||2 Aug 2011||Intel Corporation||Register set used in multithreaded parallel processor architecture|
|US8001364||23 Oct 2009||16 Aug 2011||Intel Corporation||Dynamically migrating channels|
|US8001549 *||26 Apr 2007||16 Aug 2011||Panasonic Corporation||Multithreaded computer system and multithread execution control method|
|US8024719||3 Nov 2008||20 Sep 2011||Advanced Micro Devices, Inc.||Bounded hash table sorting in a dynamic program profiling system|
|US8024735 *||14 Jun 2002||20 Sep 2011||Intel Corporation||Method and apparatus for ensuring fairness and forward progress when executing multiple threads of execution|
|US8037250||9 Dec 2004||11 Oct 2011||Oracle America, Inc.||Arbitrating cache misses in a multithreaded/multi-core processor|
|US8042116 *||1 Jun 2005||18 Oct 2011||Panasonic Corporation||Task switching based on the execution control information held in register groups|
|US8074051||1 Apr 2005||6 Dec 2011||Aspen Acquisition Corporation||Multithreaded processor with multiple concurrent pipelines per thread|
|US8079031||21 Oct 2005||13 Dec 2011||Intel Corporation||Method, apparatus, and a system for dynamically configuring a prefetcher based on a thread specific latency metric|
|US8095778||30 Jun 2004||10 Jan 2012||Open Computing Trust I & II||Method and system for sharing functional units of a multithreaded processor|
|US8108862 *||12 Dec 2005||31 Jan 2012||Infineon Technologies Ag||Out-of-order thread scheduling based on processor idle time thresholds|
|US8141098||16 Jan 2009||20 Mar 2012||International Business Machines Corporation||Context switch data prefetching in multithreaded computer|
|US8145884||23 Oct 2009||27 Mar 2012||Mips Technologies, Inc.||Apparatus, method and instruction for initiation of concurrent instruction streams in a multithreading microprocessor|
|US8225034||30 Jun 2004||17 Jul 2012||Oracle America, Inc.||Hybrid instruction buffer|
|US8266620||26 Oct 2010||11 Sep 2012||Mips Technologies, Inc.|
|US8285973||4 Aug 2008||9 Oct 2012||International Business Machines Corporation||Thread completion rate controlled scheduling|
|US8296552||29 Jun 2011||23 Oct 2012||Intel Corporation||Dynamically migrating channels|
|US8341604||15 Nov 2006||25 Dec 2012||Qualcomm Incorporated||Embedded trace macrocell for enhanced digital signal processor debugging operations|
|US8370806||15 Nov 2006||5 Feb 2013||Qualcomm Incorporated||Non-intrusive, thread-selective, debugging method and system for a multi-thread digital signal processor|
|US8380966||15 Nov 2006||19 Feb 2013||Qualcomm Incorporated||Method and system for instruction stuffing operations during non-intrusive digital signal processor debugging|
|US8438369||8 Nov 2010||7 May 2013||Intel Corporation||Providing thread fairness by biasing selection away from a stalling thread using a stall-cycle counter in a hyper-threaded microprocessor|
|US8447959||10 Aug 2010||21 May 2013||Fujitsu Limited||Multithread processor and method of controlling multithread processor|
|US8478948||4 Dec 2008||2 Jul 2013||Oracle America, Inc.||Method and system for efficient tracing and profiling of memory accesses during program execution|
|US8484516||11 Apr 2007||9 Jul 2013||Qualcomm Incorporated||Inter-thread trace alignment method and system for a multi-threaded processor|
|US8521993 *||9 Apr 2007||27 Aug 2013||Intel Corporation||Providing thread fairness by biasing selection away from a stalling thread using a stall-cycle counter in a hyper-threaded microprocessor|
|US8533530||15 Nov 2006||10 Sep 2013||Qualcomm Incorporated||Method and system for trusted/untrusted digital signal processor debugging operations|
|US8560814 *||4 May 2010||15 Oct 2013||Oracle International Corporation||Thread fairness on a multi-threaded processor with multi-cycle cryptographic operations|
|US8612986||20 Jun 2012||17 Dec 2013||International Business Machines Corporation||Computer program product for scheduling ready threads in a multiprocessor computer based on an interrupt mask flag value associated with a thread and a current processor priority register value|
|US8694976 *||19 Dec 2003||8 Apr 2014||Intel Corporation||Sleep state mechanism for virtual multithreading|
|US8713444||8 Mar 2011||29 Apr 2014||Microsoft Corporation||System and method for browsing tabbed-heterogeneous windows|
|US8762688||27 Oct 2011||24 Jun 2014||Qualcomm Incorporated||Multithreaded processor with multiple concurrent pipelines per thread|
|US8769246||14 Feb 2011||1 Jul 2014||Open Computing Trust I & II||Mechanism for selecting instructions for execution in a multithreaded processor|
|US8799872||27 Jun 2010||5 Aug 2014||International Business Machines Corporation||Sampling with sample pacing|
|US8799904||21 Jan 2011||5 Aug 2014||International Business Machines Corporation||Scalable system call stack sampling|
|US8843684||11 Jun 2010||23 Sep 2014||International Business Machines Corporation||Performing call stack sampling by setting affinity of target thread to a current process to prevent target thread migration|
|US8892849||15 Oct 2009||18 Nov 2014||Qualcomm Incorporated||Multithreaded processor with multiple concurrent pipelines per thread|
|US8918627 *||15 Oct 2009||23 Dec 2014||Qualcomm Incorporated||Multithreaded processor with multiple concurrent pipelines per thread|
|US8959315||15 Oct 2009||17 Feb 2015||Qualcomm Incorporated||Multithreaded processor with multiple concurrent pipelines per thread|
|US9015504||6 Jan 2011||21 Apr 2015||Texas Instruments Incorporated||Managing power of thread pipelines according to clock frequency and voltage specified in thread registers|
|US9032404||20 Dec 2005||12 May 2015||Mips Technologies, Inc.||Preemptive multitasking employing software emulation of directed exceptions in a multithreading processor|
|US9047116||24 Apr 2008||2 Jun 2015||International Business Machines Corporation||Context switch data prefetching in multithreaded computer|
|US9176783||24 May 2010||3 Nov 2015||International Business Machines Corporation||Idle transitions sampling with execution context|
|US9218185||27 Mar 2014||22 Dec 2015||International Business Machines Corporation||Multithreading capability information retrieval|
|US9348644 *||8 Oct 2014||24 May 2016||International Business Machines Corporation||Application-level dispatcher control of application-level pseudo threads and operating system threads|
|US9354883||27 Mar 2014||31 May 2016||International Business Machines Corporation||Dynamic enablement of multithreading|
|US9389862||27 Mar 2014||12 Jul 2016||International Business Machines Corporation||Thread context restoration in a multithreading computer system|
|US9389869||6 Jan 2011||12 Jul 2016||Texas Instruments Incorporated||Multithreaded processor with plurality of scoreboards each issuing to plurality of pipelines|
|US9396018 *||10 Oct 2014||19 Jul 2016||Salesforce.Com, Inc.||Low latency architecture with directory service for integration of transactional data system with analytical data structures|
|US9417876||27 Mar 2014||16 Aug 2016||International Business Machines Corporation||Thread context restoration in a multithreading computer system|
|US9418005||22 Sep 2008||16 Aug 2016||International Business Machines Corporation||Managing garbage collection in a data processing system|
|US9454372||6 Aug 2015||27 Sep 2016||International Business Machines Corporation||Thread context restoration in a multithreading computer system|
|US9459875||6 Aug 2015||4 Oct 2016||International Business Machines Corporation||Dynamic enablement of multithreading|
|US9524191||8 Nov 2010||20 Dec 2016||Intel Corporation||Apparatus including a stall counter to bias processing element selection, and masks to allocate reservation unit entries to one or more processing elements|
|US9594660||27 Mar 2014||14 Mar 2017||International Business Machines Corporation||Multithreading computer system and program product for executing a query instruction for idle time accumulation among cores|
|US9594661||18 Aug 2015||14 Mar 2017||International Business Machines Corporation||Method for executing a query instruction for idle time accumulation among cores in a multithreading computer system|
|US9626079||4 Apr 2014||18 Apr 2017||Microsoft Technology Licensing, Llc||System and method for browsing tabbed-heterogeneous windows|
|US9804846||27 Mar 2014||31 Oct 2017||International Business Machines Corporation||Thread context preservation in a multithreading computer system|
|US9804847||6 Aug 2015||31 Oct 2017||International Business Machines Corporation||Thread context preservation in a multithreading computer system|
|US9817693 *||14 Mar 2014||14 Nov 2017||International Business Machines Corporation||Coherence protocol augmentation to indicate transaction status|
|US20020095614 *||8 Mar 2002||18 Jul 2002||Intel Corporation||Method and apparatus for disabling a clock signal within a multithreaded processor|
|US20030005251 *||11 Jan 2002||2 Jan 2003||Wilson Kenneth M.||Managing latencies in accessing memory of computer systems|
|US20030018684 *||17 Jul 2002||23 Jan 2003||Nec Corporation||Multi-thread execution method and parallel processor system|
|US20030135781 *||17 Jan 2002||17 Jul 2003||International Business Machines Corporation||Method and system for error detection in a managed application environment|
|US20030177229 *||21 Feb 2003||18 Sep 2003||Fujitsu Limited||Microcomputer, bus control circuit, and data access method for a microcomputer|
|US20030233394 *||14 Jun 2002||18 Dec 2003||Rudd Kevin W.||Method and apparatus for ensuring fairness and forward progress when executing multiple threads of execution|
|US20040015967 *||3 May 2002||22 Jan 2004||Dale Morris||Method and system for application managed context switching|
|US20040060052 *||19 Sep 2002||25 Mar 2004||International Business Machines Corporation||Method and apparatus for implementing thread replacement for optimal performance in a two-tiered multithreading structure|
|US20040064818 *||16 Sep 2003||1 Apr 2004||Alverson Gail A.||Deferred task swapping in a multithreaded environment|
|US20040073910 *||1 Apr 2003||15 Apr 2004||Erdem Hokenek||Method and apparatus for high speed cross-thread interrupts in a multithreaded processor|
|US20040128476 *||26 Dec 2002||1 Jul 2004||Robert Nuckolls||Scheme to simplify instruction buffer logic supporting multiple strands|
|US20040128488 *||26 Dec 2002||1 Jul 2004||Thimmannagari Chandra M. R.||Strand switching algorithm to avoid strand starvation|
|US20040139093 *||23 Oct 2003||15 Jul 2004||International Business Machines Corporation||Exclusion control|
|US20040154021 *||30 Jan 2003||5 Aug 2004||Vasudevan Sangili||Apparatus and method to minimize blocking overhead in upcall based MxN threads|
|US20040215984 *||24 Apr 2003||28 Oct 2004||International Business Machines Corporation||Method and circuitry for managing power in a simultaneous multithread processor|
|US20040268101 *||30 Jun 2003||30 Dec 2004||Mehta Kalpesh D||Determining if a register is ready to exchange data with a processing element|
|US20050050305 *||10 Oct 2003||3 Mar 2005||Kissell Kevin D.||Integrated mechanism for suspension and deallocation of computational threads of execution in a processor|
|US20050050395 *||10 Oct 2003||3 Mar 2005||Kissell Kevin D.||Mechanisms for assuring quality of service for programs executing on a multithreaded processor|
|US20050060517 *||12 Sep 2003||17 Mar 2005||Morrow Michael W.||Switching processor threads during long latencies|
|US20050120194 *||27 Aug 2004||2 Jun 2005||Mips Technologies, Inc.||Apparatus, method, and instruction for initiation of concurrent instruction streams in a multithreading microprocessor|
|US20050125629 *||27 Aug 2004||9 Jun 2005||Mips Technologies, Inc.||Mechanisms for dynamic configuration of virtual processor resources|
|US20050125795 *||27 Aug 2004||9 Jun 2005||Mips Technologies, Inc.||Integrated mechanism for suspension and deallocation of computational threads of execution in a processor|
|US20050138627 *||18 Dec 2003||23 Jun 2005||International Business Machines Corporation||Context switch data prefetching in multithreaded computer|
|US20050138628 *||18 Dec 2003||23 Jun 2005||International Business Machines Corporation||Context switch instruction prefetching in multithreaded computer|
|US20050138629 *||19 Dec 2003||23 Jun 2005||Samra Nicholas G.||Sleep state mechanism for virtual multithreading|
|US20050149700 *||19 Dec 2003||7 Jul 2005||Samra Nicholas G.||Virtual multithreading translation mechanism including retrofit capability|
|US20050172102 *||2 Feb 2005||4 Aug 2005||Takeshi Inuo||Array-type computer processor|
|US20050172103 *||3 Feb 2005||4 Aug 2005||Takeshi Inuo||Array-type computer processor|
|US20050240936 *||27 Aug 2004||27 Oct 2005||Mips Technologies, Inc.||Apparatus, method, and instruction for software management of multiple computational contexts in a multithreaded microprocessor|
|US20050246461 *||29 Apr 2004||3 Nov 2005||International Business Machines Corporation||Scheduling threads in a multi-processor computer|
|US20050251613 *||30 Sep 2004||10 Nov 2005||Mips Technologies, Inc., A Delaware Corporation||Synchronized storage providing multiple synchronization semantics|
|US20050251639 *||30 Sep 2004||10 Nov 2005||Mips Technologies, Inc. A Delaware Corporation||Smart memory based synchronization controller for a multi-threaded multiprocessor SoC|
|US20060004988 *||30 Jun 2004||5 Jan 2006||Jordan Paul J||Single bit control of threads in a multithreaded multicore processor|
|US20060004989 *||30 Jun 2004||5 Jan 2006||Sun Microsystems, Inc.||Mechanism for selecting instructions for execution in a multithreaded processor|
|US20060004995 *||30 Jun 2004||5 Jan 2006||Sun Microsystems, Inc.||Apparatus and method for fine-grained multithreading in a multipipelined processor core|
|US20060005051 *||30 Jun 2004||5 Jan 2006||Sun Microsystems, Inc.||Thread-based clock enabling in a multi-threaded processor|
|US20060020831 *||30 Jun 2004||26 Jan 2006||Sun Microsystems, Inc.||Method and appratus for power throttling in a multi-thread processor|
|US20060059485 *||13 Sep 2004||16 Mar 2006||Onufryk Peter Z||System and method of scheduling computing threads|
|US20060059487 *||14 Jan 2005||16 Mar 2006||Integrated Device Technology, Inc.||Scheduler for a multiprocessing computing system|
|US20060095729 *||1 Apr 2005||4 May 2006||Erdem Hokenek||Multithreaded processor with multiple concurrent pipelines per thread|
|US20060143616 *||22 Dec 2005||29 Jun 2006||Oki Electric Industry Co., Ltd.||System and method for performing multi-task processing|
|US20060156306 *||12 Dec 2005||13 Jul 2006||Infineon Technologies Ag||Thread scheduling method, and device to be used with a thread scheduling method|
|US20060161921 *||20 Dec 2005||20 Jul 2006||Mips Technologies, Inc.||Preemptive multitasking employing software emulation of directed exceptions in a multithreading processor|
|US20060184537 *||15 Feb 2005||17 Aug 2006||Microsoft Corporation||System and method for browsing tabbed-heterogeneous windows|
|US20060190945 *||11 Jan 2006||24 Aug 2006||Mips Technologies, Inc.||Symmetric multiprocessor operating system for execution on non-independent lightweight thread context|
|US20060190946 *||11 Jan 2006||24 Aug 2006||Mips Technologies, Inc.||Symmetric multiprocessor operating system for execution on non-independent lightweight thread context|
|US20060195683 *||11 Jan 2006||31 Aug 2006||Mips Technologies, Inc.|
|US20060206902 *||14 Mar 2005||14 Sep 2006||Sujat Jamil||Variable interleaved multithreaded processor method and system|
|US20060212874 *||12 Dec 2003||21 Sep 2006||Johnson Erik J||Inserting instructions|
|US20060218559 *||23 Mar 2005||28 Sep 2006||Muhammad Ahmed||Method and system for variable thread allocation and switching in a multithreaded processor|
|US20070043935 *||11 Jan 2006||22 Feb 2007||Mips Technologies, Inc.|
|US20070044104 *||18 Aug 2005||22 Feb 2007||International Business Machines Corporation||Adaptive scheduling and management of work processing in a target context in resource contention|
|US20070044105 *||11 Jan 2006||22 Feb 2007||Mips Technologies, Inc.|
|US20070044106 *||11 Jan 2006||22 Feb 2007||Mips Technologies, Inc.|
|US20070050595 *||25 Aug 2005||1 Mar 2007||Kimming So||System and method for synchronizing translation lookaside buffer access in a multithread processor|
|US20070079020 *||30 Sep 2005||5 Apr 2007||Gautham Chinya||Dynamically migrating channels|
|US20070088680 *||14 Oct 2005||19 Apr 2007||Microsoft Corporation||Simultaneously spawning multiple searches across multiple providers|
|US20070094453 *||21 Oct 2005||26 Apr 2007||Santhanakrishnan Geeyarpuram N||Method, apparatus, and a system for a software configurable prefetcher|
|US20070106887 *||23 Dec 2006||10 May 2007||Mips Technologies, Inc.|
|US20070106988 *||23 Dec 2006||10 May 2007||Mips Technologies, Inc.|
|US20070106989 *||23 Dec 2006||10 May 2007||Mips Technologies, Inc.|
|US20070106990 *||23 Dec 2006||10 May 2007||Mips Technologies, Inc.|
|US20070186028 *||30 Sep 2004||9 Aug 2007||Mips Technologies, Inc.||Synchronized storage providing multiple synchronization semantics|
|US20070204137 *||23 Aug 2006||30 Aug 2007||Texas Instruments Incorporated||Multi-threading processors, integrated circuit devices, systems, and processes of operation and manufacture|
|US20070226454 *||1 Mar 2006||27 Sep 2007||Marius Stoian||Highly scalable MIMD machine for java and .net processing|
|US20070234009 *||6 Jun 2007||4 Oct 2007||Intel Corporation||Processor having a dedicated hash unit integrated within|
|US20070260853 *||11 Jul 2007||8 Nov 2007||Morrow Michael W||Switching processor threads during long latencies|
|US20070266387 *||26 Apr 2007||15 Nov 2007||Matsushita Electric Industrial Co., Ltd.||Multithreaded computer system and multithread execution control method|
|US20080114972 *||15 Nov 2006||15 May 2008||Lucian Codrescu||Method and system for instruction stuffing operations during non-intrusive digital signal processor debugging|
|US20080115011 *||15 Nov 2006||15 May 2008||Lucian Codrescu||Method and system for trusted/untrusted digital signal processor debugging operations|
|US20080115115 *||15 Nov 2006||15 May 2008||Lucian Codrescu||Embedded trace macrocell for enhanced digital signal processor debugging operations|
|US20080140998 *||3 Dec 2007||12 Jun 2008||Mips Technologies, Inc.|
|US20080177979 *||28 Mar 2008||24 Jul 2008||Gheorghe Stefan||Hardware multi-core processor optimized for object oriented computing|
|US20080178183 *||25 Mar 2008||24 Jul 2008||International Business Machines Corporation||Scheduling Threads In A Multi-Processor Computer|
|US20080201529 *||24 Apr 2008||21 Aug 2008||International Business Machines Corporation||Context switch data prefetching in multithreaded computer|
|US20080201565 *||24 Apr 2008||21 Aug 2008||International Business Machines Corporation||Context switch data prefetching in multithreaded computer|
|US20080243887 *||4 Apr 2008||2 Oct 2008||International Business Machines Corp.||Exclusion control|
|US20080250233 *||9 Apr 2007||9 Oct 2008||Morris Marden||Providing thread fairness in a hyper-threaded microprocessor|
|US20080256396 *||11 Apr 2007||16 Oct 2008||Louis Achille Giannini||Inter-thread trace alignment method and system for a multi-threaded processor|
|US20080256542 *||1 Jun 2005||16 Oct 2008||Kazushi Kurata||Processor|
|US20080271042 *||24 Apr 2007||30 Oct 2008||Microsoft Corporation||Testing multi-thread software using prioritized context switch limits|
|US20080313652 *||15 Jun 2007||18 Dec 2008||Microsoft Corporation||Notifying user mode scheduler of blocking events|
|US20090125913 *||16 Jan 2009||14 May 2009||International Business Machines Corporation||Context switch data prefetching in multithreaded computer|
|US20100017583 *||15 Jul 2008||21 Jan 2010||International Business Machines Corporation||Call Stack Sampling for a Multi-Processor System|
|US20100115494 *||3 Nov 2008||6 May 2010||Gorton Jr Richard C||System for dynamic program profiling|
|US20100122068 *||15 Oct 2009||13 May 2010||Erdem Hokenek||Multithreaded processor with multiple concurrent pipelines per thread|
|US20100125722 *||20 Nov 2008||20 May 2010||International Business Machines Corporation||Multithreaded Processing Unit With Thread Pair Context Caching|
|US20100146220 *||4 Dec 2008||10 Jun 2010||Panchenko Maksim V||Efficient program instrumentation for memory profiling|
|US20100199073 *||15 Oct 2009||5 Aug 2010||Erdem Hokenek||Multithreaded processor with multiple concurrent pipelines per thread|
|US20100199075 *||15 Oct 2009||5 Aug 2010||Erdem Hokenek||Multithreaded processor with multiple concurrent pipelines per thread|
|US20100325396 *||10 Aug 2010||23 Dec 2010||Fujitsu Limited||Multithread processor and register control method|
|US20110040956 *||26 Oct 2010||17 Feb 2011||Mips Technologies, Inc.||Symmetric Multiprocessor Operating System for Execution On Non-Independent Lightweight Thread Contexts|
|US20110055524 *||8 Nov 2010||3 Mar 2011||Morris Marden||Providing thread fairness in a hyper-threaded microprocessor|
|US20110055525 *||8 Nov 2010||3 Mar 2011||Morris Marden||Providing thread fairness in a hyper-threaded microprocessor|
|US20110099355 *||6 Jan 2011||28 Apr 2011||Texas Instruments Incorporated||Multi-threading processors, integrated circuit devices, systems, and processes of operation and manufacture|
|US20110099393 *||6 Jan 2011||28 Apr 2011||Texas Instruments Incorporated|
|US20110161828 *||8 Mar 2011||30 Jun 2011||Microsoft Corporation||System and Method for Browsing Tabbed-Heterogeneous Windows|
|US20110276783 *||4 May 2010||10 Nov 2011||Golla Robert T||Thread fairness on a multi-threaded processor with multi-cycle cryptographic operations|
|US20150261564 *||30 Sep 2014||17 Sep 2015||International Business Machines Corporation||Coherence protocol augmentation to indicate transaction status|
|US20150261676 *||14 Mar 2014||17 Sep 2015||International Business Machines Corporation||Coherence protocol augmentation to indicate transaction status|
|US20160092235 *||26 Sep 2014||31 Mar 2016||Rekai Gonzalez-Alberquilla||Method and apparatus for improved thread selection|
|CN101013378B||30 Sep 2006||11 Aug 2010||英特尔公司||Dynamically migrating channels|
|WO2008157455A2 *||15 Jun 2008||24 Dec 2008||Microsoft Corporation||Notifying user mode scheduler of blocking events|
|WO2008157455A3 *||15 Jun 2008||5 Mar 2009||Microsoft Corp||Notifying user mode scheduler of blocking events|
|U.S. Classification||712/228, 712/E09.053, 719/318, 718/108|
|International Classification||G06F9/46, G06F9/48, G06F9/38, G06F12/08|
|Cooperative Classification||G06F9/4843, G06F9/3851|
|European Classification||G06F9/48C4, G06F9/38E4|
|23 Oct 1997||AS||Assignment|
Owner name: INTERNATIONAL BUSINESS MACHINES CORPORATION, NEW Y
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:BORKENHAGEN, JOHN MICHAEL;EICKEMEYER, RICHARD JAMES;FLYNN, WILLIAM THOMAS;AND OTHERS;REEL/FRAME:008801/0347;SIGNING DATES FROM 19971001 TO 19971006
|13 Jul 2007||FPAY||Fee payment|
Year of fee payment: 4
|15 Jul 2011||FPAY||Fee payment|
Year of fee payment: 8
|30 Jun 2015||FPAY||Fee payment|
Year of fee payment: 12